[{"data":1,"prerenderedAt":2180},["ShallowReactive",2],{"book-hierarchy":3,"content-query-0gLdGIcrva":1073},[4,262,999],{"key":5,"slug":5,"title":6,"cleanTitle":6,"numberedTitle":7,"description":8,"path":9,"basePath":10,"filePath":11,"pageType":12,"hidden":13,"effectiveHidden":13,"order":14,"depth":14,"number":15,"ownAudience":16,"audiences":17,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":22},"01_programmierung","Programmierung","I Programmierung","Grundlagen der Programmierung.","/01_programmierung/de","/01_programmierung","content/01_Programmierung/de.md","chapter",false,1,"I","Sek I",[16],"cover-de-light.png","cover-en-light.png","cover-de-dark.png","cover-en-dark.png",[23,37,250],{"key":24,"slug":25,"title":26,"cleanTitle":26,"numberedTitle":27,"description":28,"path":29,"basePath":30,"filePath":31,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":33,"number":34,"ownAudience":16,"audiences":35,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":36},"01_programmierung/01_python","01_python","Python","I.1 Python","Einführung in die Programmiersprache Python.","/01_programmierung/01_python/de","/01_programmierung/01_python","content/01_Programmierung/01_Python/de.md","content",2,"I.1",[16],[],{"key":38,"slug":39,"title":40,"cleanTitle":40,"numberedTitle":41,"description":42,"path":43,"basePath":44,"filePath":45,"pageType":12,"hidden":13,"effectiveHidden":13,"order":33,"depth":33,"number":46,"ownAudience":16,"audiences":47,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":48},"01_programmierung/02_java","02_java","Java","I.2 Java","Objektorientierte Programmierung mit Java: von OOP-Grundlagen bis zu Projekten.","/01_programmierung/02_java/de","/01_programmierung/02_java","content/01_Programmierung/02_Java/de.md","I.2",[16],[49,139,190,202,214],{"key":50,"slug":51,"title":52,"cleanTitle":52,"numberedTitle":53,"description":54,"path":55,"basePath":56,"filePath":57,"pageType":58,"hidden":13,"effectiveHidden":13,"order":14,"depth":59,"number":60,"ownAudience":16,"audiences":61,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":63},"01_programmierung/02_java/01_oop_teil_i","01_oop_teil_i","OOP Teil I","I.2.1 OOP Teil I","Grundlagen der objektorientierten Programmierung mit Java.","/01_programmierung/02_java/01_oop_teil_i/de","/01_programmierung/02_java/01_oop_teil_i","content/01_Programmierung/02_Java/01_OOP_Teil_I/de.md","overview",3,"I.2.1",[16],"",[64,78,90,102,114,126],{"key":65,"slug":66,"title":67,"cleanTitle":67,"numberedTitle":68,"description":69,"path":70,"basePath":71,"filePath":72,"pageType":32,"hidden":13,"effectiveHidden":13,"order":73,"depth":74,"number":75,"ownAudience":16,"audiences":76,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":77},"01_programmierung/02_java/01_oop_teil_i/00_uebersicht_installation","00_uebersicht_installation","Uebersicht und Installation","I.2.1.1 Uebersicht und Installation","Programmierparadigmen, Java-Grundidee und Einrichtung der Entwicklungsumgebung.","/01_programmierung/02_java/01_oop_teil_i/00_uebersicht_installation/de","/01_programmierung/02_java/01_oop_teil_i/00_uebersicht_installation","content/01_Programmierung/02_Java/01_OOP_Teil_I/00_Uebersicht_Installation/de.md",0,4,"I.2.1.1",[16],[],{"key":79,"slug":80,"title":81,"cleanTitle":81,"numberedTitle":82,"description":83,"path":84,"basePath":85,"filePath":86,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":74,"number":87,"ownAudience":16,"audiences":88,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":89},"01_programmierung/02_java/01_oop_teil_i/01_klassen_und_objekte","01_klassen_und_objekte","Klassen und Objekte","I.2.1.2 Klassen und Objekte","Modellierung mit Klassen, Attribute, Methoden und erste Objektinstanzen.","/01_programmierung/02_java/01_oop_teil_i/01_klassen_und_objekte/de","/01_programmierung/02_java/01_oop_teil_i/01_klassen_und_objekte","content/01_Programmierung/02_Java/01_OOP_Teil_I/01_Klassen_und_Objekte/de.md","I.2.1.2",[16],[],{"key":91,"slug":92,"title":93,"cleanTitle":93,"numberedTitle":94,"description":95,"path":96,"basePath":97,"filePath":98,"pageType":32,"hidden":13,"effectiveHidden":13,"order":33,"depth":74,"number":99,"ownAudience":16,"audiences":100,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":101},"01_programmierung/02_java/01_oop_teil_i/02_datentypen_methoden_parameter","02_datentypen_methoden_parameter","Datentypen, Methoden und Parameter","I.2.1.3 Datentypen, Methoden und Parameter","Primitive Typen, lokale Variablen, Methodensignaturen und Rueckgabewerte.","/01_programmierung/02_java/01_oop_teil_i/02_datentypen_methoden_parameter/de","/01_programmierung/02_java/01_oop_teil_i/02_datentypen_methoden_parameter","content/01_Programmierung/02_Java/01_OOP_Teil_I/02_Datentypen_Methoden_Parameter/de.md","I.2.1.3",[16],[],{"key":103,"slug":104,"title":105,"cleanTitle":105,"numberedTitle":106,"description":107,"path":108,"basePath":109,"filePath":110,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":74,"number":111,"ownAudience":16,"audiences":112,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":113},"01_programmierung/02_java/01_oop_teil_i/03_attribute_kapselung_setter_getter","03_attribute_kapselung_setter_getter","Attribute, Kapselung, Setter und Getter","I.2.1.4 Attribute, Kapselung, Setter und Getter","Saubere Objektzustaende durch Zugriffskontrolle und kontrollierte Schnittstellen.","/01_programmierung/02_java/01_oop_teil_i/03_attribute_kapselung_setter_getter/de","/01_programmierung/02_java/01_oop_teil_i/03_attribute_kapselung_setter_getter","content/01_Programmierung/02_Java/01_OOP_Teil_I/03_Attribute_Kapselung_Setter_Getter/de.md","I.2.1.4",[16],[],{"key":115,"slug":116,"title":117,"cleanTitle":117,"numberedTitle":118,"description":119,"path":120,"basePath":121,"filePath":122,"pageType":32,"hidden":13,"effectiveHidden":13,"order":74,"depth":74,"number":123,"ownAudience":16,"audiences":124,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":125},"01_programmierung/02_java/01_oop_teil_i/04_new_konstruktoren_ueberladen","04_new_konstruktoren_ueberladen","new, Konstruktoren und Ueberladen","I.2.1.5 new, Konstruktoren und Ueberladen","Objekte gezielt erzeugen und flexible Schnittstellen durch Ueberladen entwerfen.","/01_programmierung/02_java/01_oop_teil_i/04_new_konstruktoren_ueberladen/de","/01_programmierung/02_java/01_oop_teil_i/04_new_konstruktoren_ueberladen","content/01_Programmierung/02_Java/01_OOP_Teil_I/04_New_Konstruktoren_Ueberladen/de.md","I.2.1.5",[16],[],{"key":127,"slug":128,"title":129,"cleanTitle":129,"numberedTitle":130,"description":131,"path":132,"basePath":133,"filePath":134,"pageType":32,"hidden":13,"effectiveHidden":13,"order":135,"depth":74,"number":136,"ownAudience":16,"audiences":137,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":138},"01_programmierung/02_java/01_oop_teil_i/05_vererbung_abstrakte_klassen","05_vererbung_abstrakte_klassen","Vererbung und abstrakte Klassen","I.2.1.6 Vererbung und abstrakte Klassen","Code wiederverwenden, Spezialisierungen bilden und gemeinsame Schnittstellen sichern.","/01_programmierung/02_java/01_oop_teil_i/05_vererbung_abstrakte_klassen/de","/01_programmierung/02_java/01_oop_teil_i/05_vererbung_abstrakte_klassen","content/01_Programmierung/02_Java/01_OOP_Teil_I/05_Vererbung_Abstrakte_Klassen/de.md",5,"I.2.1.6",[16],[],{"key":140,"slug":141,"title":142,"cleanTitle":142,"numberedTitle":143,"description":144,"path":145,"basePath":146,"filePath":147,"pageType":58,"hidden":13,"effectiveHidden":13,"order":33,"depth":59,"number":148,"ownAudience":16,"audiences":149,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":150},"01_programmierung/02_java/02_schleifen_arrays_statischemethoden","02_schleifen_arrays_statischemethoden","Schleifen, Arrays und statische Methoden","I.2.2 Schleifen, Arrays und statische Methoden","Kontrollstrukturen, Array-Datenstrukturen und der Einsatz von static.","/01_programmierung/02_java/02_schleifen_arrays_statischemethoden/de","/01_programmierung/02_java/02_schleifen_arrays_statischemethoden","content/01_Programmierung/02_Java/02_Schleifen_Arrays_StatischeMethoden/de.md","I.2.2",[16],[151,164,177],{"key":152,"slug":153,"title":154,"cleanTitle":154,"numberedTitle":155,"description":156,"path":157,"basePath":158,"filePath":159,"pageType":32,"hidden":13,"effectiveHidden":13,"order":160,"depth":74,"number":161,"ownAudience":16,"audiences":162,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":163},"01_programmierung/02_java/02_schleifen_arrays_statischemethoden/06_statische_variablen_und_methoden","06_statische_variablen_und_methoden","Statische Variablen und Methoden","I.2.2.1 Statische Variablen und Methoden","Unterschied zwischen Instanz- und Klassenebene in Java.","/01_programmierung/02_java/02_schleifen_arrays_statischemethoden/06_statische_variablen_und_methoden/de","/01_programmierung/02_java/02_schleifen_arrays_statischemethoden/06_statische_variablen_und_methoden","content/01_Programmierung/02_Java/02_Schleifen_Arrays_StatischeMethoden/06_Statische_Variablen_und_Methoden/de.md",6,"I.2.2.1",[16],[],{"key":165,"slug":166,"title":167,"cleanTitle":167,"numberedTitle":168,"description":169,"path":170,"basePath":171,"filePath":172,"pageType":32,"hidden":13,"effectiveHidden":13,"order":173,"depth":74,"number":174,"ownAudience":16,"audiences":175,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":176},"01_programmierung/02_java/02_schleifen_arrays_statischemethoden/07_arrays","07_arrays","Arrays","I.2.2.2 Arrays","Ein- und mehrdimensionale Arrays, Initialisierung und typische Fehlerquellen.","/01_programmierung/02_java/02_schleifen_arrays_statischemethoden/07_arrays/de","/01_programmierung/02_java/02_schleifen_arrays_statischemethoden/07_arrays","content/01_Programmierung/02_Java/02_Schleifen_Arrays_StatischeMethoden/07_Arrays/de.md",7,"I.2.2.2",[16],[],{"key":178,"slug":179,"title":180,"cleanTitle":180,"numberedTitle":181,"description":182,"path":183,"basePath":184,"filePath":185,"pageType":32,"hidden":13,"effectiveHidden":13,"order":186,"depth":74,"number":187,"ownAudience":16,"audiences":188,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":189},"01_programmierung/02_java/02_schleifen_arrays_statischemethoden/08_schleifen_und_uebungen","08_schleifen_und_uebungen","Schleifen und Uebungen","I.2.2.3 Schleifen und Uebungen","for- und while-Schleifen, Summenformeln, Suche und verschachtelte Iteration.","/01_programmierung/02_java/02_schleifen_arrays_statischemethoden/08_schleifen_und_uebungen/de","/01_programmierung/02_java/02_schleifen_arrays_statischemethoden/08_schleifen_und_uebungen","content/01_Programmierung/02_Java/02_Schleifen_Arrays_StatischeMethoden/08_Schleifen_und_Uebungen/de.md",8,"I.2.2.3",[16],[],{"key":191,"slug":192,"title":193,"cleanTitle":193,"numberedTitle":194,"description":195,"path":196,"basePath":197,"filePath":198,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":59,"number":199,"ownAudience":16,"audiences":200,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":201},"01_programmierung/02_java/03_strukturierenmitmethoden","03_strukturierenmitmethoden","Strukturieren mit Methoden","I.2.3 Strukturieren mit Methoden","Lange Methoden in klar abgegrenzte Teilmethoden zerlegen.","/01_programmierung/02_java/03_strukturierenmitmethoden/de","/01_programmierung/02_java/03_strukturierenmitmethoden","content/01_Programmierung/02_Java/03_StrukturierenMitMethoden/de.md","I.2.3",[16],[],{"key":203,"slug":204,"title":205,"cleanTitle":205,"numberedTitle":206,"description":207,"path":208,"basePath":209,"filePath":210,"pageType":32,"hidden":13,"effectiveHidden":13,"order":74,"depth":59,"number":211,"ownAudience":16,"audiences":212,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":213},"01_programmierung/02_java/04_gueltigkeitundsichtbarkeit","04_gueltigkeitundsichtbarkeit","Gueltigkeit und Sichtbarkeit","I.2.4 Gueltigkeit und Sichtbarkeit","Scope, Lebensdauer und Zugriffsregeln von Variablen und Attributen.","/01_programmierung/02_java/04_gueltigkeitundsichtbarkeit/de","/01_programmierung/02_java/04_gueltigkeitundsichtbarkeit","content/01_Programmierung/02_Java/04_GueltigkeitUndSichtbarkeit/de.md","I.2.4",[16],[],{"key":215,"slug":216,"title":217,"cleanTitle":217,"numberedTitle":218,"description":219,"path":220,"basePath":221,"filePath":222,"pageType":58,"hidden":13,"effectiveHidden":13,"order":74,"depth":59,"number":223,"ownAudience":16,"audiences":224,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":225},"01_programmierung/02_java/04_oop_teil_ii","04_oop_teil_ii","OOP Teil II","I.2.5 OOP Teil II","Vertiefung zu Polymorphie, Typisierung, Interfaces und Modellierungsfragen.","/01_programmierung/02_java/04_oop_teil_ii/de","/01_programmierung/02_java/04_oop_teil_ii","content/01_Programmierung/02_Java/04_OOP_Teil_II/de.md","I.2.5",[16],[226,238],{"key":227,"slug":228,"title":229,"cleanTitle":229,"numberedTitle":230,"description":231,"path":232,"basePath":233,"filePath":234,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":74,"number":235,"ownAudience":16,"audiences":236,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":237},"01_programmierung/02_java/04_oop_teil_ii/01_uebungen","01_uebungen","Uebungen und Vertiefung","I.2.5.1 Uebungen und Vertiefung","Kovarianz, Polymorphie, statische Typisierung und erlaubte Zuweisungen.","/01_programmierung/02_java/04_oop_teil_ii/01_uebungen/de","/01_programmierung/02_java/04_oop_teil_ii/01_uebungen","content/01_Programmierung/02_Java/04_OOP_Teil_II/01_Uebungen/de.md","I.2.5.1",[16],[],{"key":239,"slug":240,"title":241,"cleanTitle":241,"numberedTitle":242,"description":243,"path":244,"basePath":245,"filePath":246,"pageType":32,"hidden":13,"effectiveHidden":13,"order":33,"depth":74,"number":247,"ownAudience":16,"audiences":248,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":249},"01_programmierung/02_java/04_oop_teil_ii/02_projekt_zahlen","02_projekt_zahlen","Projekt Zahlen","I.2.5.2 Projekt Zahlen","Natuerliche, ganze und rationale Zahlen mit Interfaces und sauberem Klassendesign.","/01_programmierung/02_java/04_oop_teil_ii/02_projekt_zahlen/de","/01_programmierung/02_java/04_oop_teil_ii/02_projekt_zahlen","content/01_Programmierung/02_Java/04_OOP_Teil_II/02_Projekt_Zahlen/de.md","I.2.5.2",[16],[],{"key":251,"slug":252,"title":253,"cleanTitle":253,"numberedTitle":254,"description":255,"path":256,"basePath":257,"filePath":258,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":33,"number":259,"ownAudience":16,"audiences":260,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":261},"01_programmierung/03_cleancode","03_cleancode","Clean Code","I.3 Clean Code","Grundlagen für sauberen und lesbaren Code.","/01_programmierung/03_cleancode/de","/01_programmierung/03_cleancode","content/01_Programmierung/03_CleanCode/de.md","I.3",[16],[],{"key":263,"slug":263,"title":264,"cleanTitle":264,"numberedTitle":265,"description":266,"path":267,"basePath":268,"filePath":269,"pageType":12,"hidden":13,"effectiveHidden":13,"order":186,"depth":14,"number":270,"ownAudience":62,"audiences":271,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":272},"08_kuenstlicheintelligenz","Künstliche Intelligenz","II Künstliche Intelligenz","Maschinelles Lernen und intelligente Systeme.","/08_kuenstlicheintelligenz/de","/08_kuenstlicheintelligenz","content/08_KuenstlicheIntelligenz/de.md","II",[16],[273,359,407,827,949],{"key":274,"slug":275,"title":276,"cleanTitle":276,"numberedTitle":277,"description":278,"path":279,"basePath":280,"filePath":281,"pageType":58,"hidden":13,"effectiveHidden":13,"order":14,"depth":33,"number":282,"ownAudience":62,"audiences":283,"coverDeLight":284,"coverEnLight":285,"coverDeDark":286,"coverEnDark":287,"children":288},"08_kuenstlicheintelligenz/01_was_ist_ki","01_was_ist_ki","Einführung in das Maschinelle Lernen","II.1 Einführung in das Maschinelle Lernen","Ein Überblick über die fundamentalen Konzepte des Maschinellen Lernens.","/08_kuenstlicheintelligenz/01_was_ist_ki/de","/08_kuenstlicheintelligenz/01_was_ist_ki","content/08_KuenstlicheIntelligenz/01_was_ist_ki/de.md","II.1",[16],"cover-de-light.webp","cover-en-light.webp","cover-de-dark.webp","cover-en-dark.webp",[289,300,312,323,335,347],{"key":290,"slug":291,"title":264,"cleanTitle":264,"numberedTitle":292,"description":293,"path":294,"basePath":295,"filePath":296,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":59,"number":297,"ownAudience":16,"audiences":298,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":299},"08_kuenstlicheintelligenz/01_was_ist_ki/01-kuenstliche-intelligenz","01-kuenstliche-intelligenz","II.1.1 Künstliche Intelligenz","Unterseite zu Künstliche Intelligenz","/08_kuenstlicheintelligenz/01_was_ist_ki/01-kuenstliche-intelligenz/de","/08_kuenstlicheintelligenz/01_was_ist_ki/01-kuenstliche-intelligenz","content/08_KuenstlicheIntelligenz/01_was_ist_ki/01-kuenstliche-intelligenz/de.md","II.1.1",[16],[],{"key":301,"slug":302,"title":303,"cleanTitle":303,"numberedTitle":304,"description":305,"path":306,"basePath":307,"filePath":308,"pageType":32,"hidden":13,"effectiveHidden":13,"order":33,"depth":59,"number":309,"ownAudience":16,"audiences":310,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":311},"08_kuenstlicheintelligenz/01_was_ist_ki/02-was-ist-ml","02-was-ist-ml","Was ist ML","II.1.2 Was ist ML","Eine Einführung in die Grundlagen des Maschinellen Lernens.","/08_kuenstlicheintelligenz/01_was_ist_ki/02-was-ist-ml/de","/08_kuenstlicheintelligenz/01_was_ist_ki/02-was-ist-ml","content/08_KuenstlicheIntelligenz/01_was_ist_ki/02-was-ist-ml/de.md","II.1.2",[16],[],{"key":313,"slug":314,"title":315,"cleanTitle":315,"numberedTitle":316,"description":62,"path":317,"basePath":318,"filePath":319,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":59,"number":320,"ownAudience":16,"audiences":321,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":322},"08_kuenstlicheintelligenz/01_was_ist_ki/03-datascience_und_datenarten","03-datascience_und_datenarten","Qualitative und quantitative Merkmale","II.1.3 Qualitative und quantitative Merkmale","/08_kuenstlicheintelligenz/01_was_ist_ki/03-datascience_und_datenarten/de","/08_kuenstlicheintelligenz/01_was_ist_ki/03-datascience_und_datenarten","content/08_KuenstlicheIntelligenz/01_was_ist_ki/03-DataScience_und_Datenarten/de.md","II.1.3",[16],[],{"key":324,"slug":325,"title":326,"cleanTitle":326,"numberedTitle":327,"description":328,"path":329,"basePath":330,"filePath":331,"pageType":32,"hidden":13,"effectiveHidden":13,"order":74,"depth":59,"number":332,"ownAudience":16,"audiences":333,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":334},"08_kuenstlicheintelligenz/01_was_ist_ki/04_featureextraction","04_featureextraction","Feature Extraction","II.1.4 Feature Extraction","Von Rohdaten wie Audioaufnahmen zu Merkmalen und Datensätzen.","/08_kuenstlicheintelligenz/01_was_ist_ki/04_featureextraction/de","/08_kuenstlicheintelligenz/01_was_ist_ki/04_featureextraction","content/08_KuenstlicheIntelligenz/01_was_ist_ki/04_featureextraction/de.md","II.1.4",[16],[],{"key":336,"slug":337,"title":338,"cleanTitle":338,"numberedTitle":339,"description":340,"path":341,"basePath":342,"filePath":343,"pageType":32,"hidden":13,"effectiveHidden":13,"order":135,"depth":59,"number":344,"ownAudience":16,"audiences":345,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":346},"08_kuenstlicheintelligenz/01_was_ist_ki/05-lernparadigmen_von_ml","05-lernparadigmen_von_ml","Arten von ML","II.1.5 Arten von ML","Ueberblick ueber Ueberwachtes, Unueberwachtes und Verstaerkendes Lernen.","/08_kuenstlicheintelligenz/01_was_ist_ki/05-lernparadigmen_von_ml/de","/08_kuenstlicheintelligenz/01_was_ist_ki/05-lernparadigmen_von_ml","content/08_KuenstlicheIntelligenz/01_was_ist_ki/05-lernparadigmen_von_ml/de.md","II.1.5",[16],[],{"key":348,"slug":349,"title":350,"cleanTitle":350,"numberedTitle":351,"description":352,"path":353,"basePath":354,"filePath":355,"pageType":32,"hidden":13,"effectiveHidden":13,"order":160,"depth":59,"number":356,"ownAudience":16,"audiences":357,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":358},"08_kuenstlicheintelligenz/01_was_ist_ki/06-ueberblick","06-ueberblick","Ueberblick","II.1.6 Ueberblick","Zusammenfassung der wichtigsten Begriffe aus dem Unterkapitel Was ist KI.","/08_kuenstlicheintelligenz/01_was_ist_ki/06-ueberblick/de","/08_kuenstlicheintelligenz/01_was_ist_ki/06-ueberblick","content/08_KuenstlicheIntelligenz/01_was_ist_ki/06-ueberblick/de.md","II.1.6",[16],[],{"key":360,"slug":361,"title":362,"cleanTitle":362,"numberedTitle":363,"description":364,"path":365,"basePath":366,"filePath":367,"pageType":12,"hidden":13,"effectiveHidden":13,"order":33,"depth":33,"number":368,"ownAudience":62,"audiences":369,"coverDeLight":284,"coverEnLight":285,"coverDeDark":286,"coverEnDark":287,"children":370},"08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele","grundlagen-und-datenbeispiele","Grundlagen und Datensatzbeispiele","II.2 Grundlagen und Datensatzbeispiele","Unterseite zu Grundlagen und Datensatzbeispiele","/08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele/de","/08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele","content/08_KuenstlicheIntelligenz/grundlagen-und-datenbeispiele/de.md","II.2",[],[371,383,395],{"key":372,"slug":373,"title":374,"cleanTitle":374,"numberedTitle":375,"description":376,"path":377,"basePath":378,"filePath":379,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":59,"number":380,"ownAudience":62,"audiences":381,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":382},"08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele/datensatzbeispiele","datensatzbeispiele","Datenarten und Datensatzbeispiele","II.2.1 Datenarten und Datensatzbeispiele","Unterseite zu Datensatzbeispiele","/08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele/datensatzbeispiele/de","/08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele/datensatzbeispiele","content/08_KuenstlicheIntelligenz/grundlagen-und-datenbeispiele/datensatzbeispiele/de.md","II.2.1",[],[],{"key":384,"slug":385,"title":386,"cleanTitle":386,"numberedTitle":387,"description":388,"path":389,"basePath":390,"filePath":391,"pageType":32,"hidden":13,"effectiveHidden":13,"order":33,"depth":59,"number":392,"ownAudience":62,"audiences":393,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":394},"08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele/dinstanzm","dinstanzm","Distanzmaße","II.2.2 Distanzmaße","Unterseite zu Distanzmaße","/08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele/dinstanzm/de","/08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele/dinstanzm","content/08_KuenstlicheIntelligenz/grundlagen-und-datenbeispiele/dinstanzm/de.md","II.2.2",[],[],{"key":396,"slug":397,"title":398,"cleanTitle":398,"numberedTitle":399,"description":400,"path":401,"basePath":402,"filePath":403,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":59,"number":404,"ownAudience":62,"audiences":405,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":406},"08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele/suche","suche","Suche, Precision und Recall","II.2.3 Suche, Precision und Recall","Unterseite zu Suche, Precision und Recall","/08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele/suche/de","/08_kuenstlicheintelligenz/grundlagen-und-datenbeispiele/suche","content/08_KuenstlicheIntelligenz/grundlagen-und-datenbeispiele/suche/de.md","II.2.3",[],[],{"key":408,"slug":409,"title":410,"cleanTitle":410,"numberedTitle":411,"description":412,"path":413,"basePath":414,"filePath":415,"pageType":12,"hidden":13,"effectiveHidden":13,"order":59,"depth":33,"number":416,"ownAudience":62,"audiences":417,"coverDeLight":284,"coverEnLight":285,"coverDeDark":286,"coverEnDark":287,"children":418},"08_kuenstlicheintelligenz/02_supervisedlearning","02_supervisedlearning","Überwachtes Lernen","II.3 Überwachtes Lernen","Unterseite zu 02_supervisedlearning","/08_kuenstlicheintelligenz/02_supervisedlearning/de","/08_kuenstlicheintelligenz/02_supervisedlearning","content/08_KuenstlicheIntelligenz/02_supervisedlearning/de.md","II.3",[16],[419,515,527,563,575,815],{"key":420,"slug":421,"title":422,"cleanTitle":422,"numberedTitle":423,"description":424,"path":425,"basePath":426,"filePath":427,"pageType":58,"hidden":13,"effectiveHidden":13,"order":14,"depth":59,"number":428,"ownAudience":62,"audiences":429,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":430},"08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl","01_grundlagen_des_sl","Grundlagen des Supervised Learnings","II.3.1 Grundlagen des Supervised Learnings","Unterseite zu Grundlagen des Supervised Learnings","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/de","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl","content/08_KuenstlicheIntelligenz/02_supervisedlearning/01_grundlagen_des_sl/de.md","II.3.1",[16],[431,443,455,467,479,491,503],{"key":432,"slug":433,"title":434,"cleanTitle":434,"numberedTitle":435,"description":436,"path":437,"basePath":438,"filePath":439,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":74,"number":440,"ownAudience":62,"audiences":441,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":442},"08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/02_klassifikation_mit_entscheidungsbaumen","02_klassifikation_mit_entscheidungsbaumen","Klassifikation mit Entscheidungsbäumen","II.3.1.1 Klassifikation mit Entscheidungsbäumen","Unterseite zu Klassifikation mit Entscheidungsbäumen","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/02_klassifikation_mit_entscheidungsbaumen/de","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/02_klassifikation_mit_entscheidungsbaumen","content/08_KuenstlicheIntelligenz/02_supervisedlearning/01_grundlagen_des_sl/02_klassifikation_mit_entscheidungsbaumen/de.md","II.3.1.1",[],[],{"key":444,"slug":445,"title":446,"cleanTitle":446,"numberedTitle":447,"description":448,"path":449,"basePath":450,"filePath":451,"pageType":32,"hidden":13,"effectiveHidden":13,"order":33,"depth":74,"number":452,"ownAudience":62,"audiences":453,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":454},"08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/03_regression_mit_entscheidungsbaemen","03_regression_mit_entscheidungsbaemen","Regression mit Entscheidungsbäumen","II.3.1.2 Regression mit Entscheidungsbäumen","Unterseite zu Regression mit Entscheidungsbäumen","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/03_regression_mit_entscheidungsbaemen/de","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/03_regression_mit_entscheidungsbaemen","content/08_KuenstlicheIntelligenz/02_supervisedlearning/01_grundlagen_des_sl/03_regression_mit_entscheidungsbaemen/de.md","II.3.1.2",[],[],{"key":456,"slug":457,"title":458,"cleanTitle":458,"numberedTitle":459,"description":460,"path":461,"basePath":462,"filePath":463,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":74,"number":464,"ownAudience":62,"audiences":465,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":466},"08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/01_problemstellungen_des_sl","01_problemstellungen_des_sl","Aufgabenstellungen des Überwachten Lernens","II.3.1.3 Aufgabenstellungen des Überwachten Lernens","Unterseite zu Aufgabenstellungen des Überwachten Lernens","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/01_problemstellungen_des_sl/de","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/01_problemstellungen_des_sl","content/08_KuenstlicheIntelligenz/02_supervisedlearning/01_grundlagen_des_sl/01_problemstellungen_des_sl/de.md","II.3.1.3",[],[],{"key":468,"slug":469,"title":470,"cleanTitle":470,"numberedTitle":471,"description":472,"path":473,"basePath":474,"filePath":475,"pageType":32,"hidden":13,"effectiveHidden":13,"order":74,"depth":74,"number":476,"ownAudience":62,"audiences":477,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":478},"08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/04_beurteilung_von_klassifikatoren","04_beurteilung_von_klassifikatoren","Beurteilung von Klassifikatoren","II.3.1.4 Beurteilung von Klassifikatoren","Unterseite zu 04_beurteilung_von_klassifikatoren","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/04_beurteilung_von_klassifikatoren/de","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/04_beurteilung_von_klassifikatoren","content/08_KuenstlicheIntelligenz/02_supervisedlearning/01_grundlagen_des_sl/04_beurteilung_von_klassifikatoren/de.md","II.3.1.4",[],[],{"key":480,"slug":481,"title":482,"cleanTitle":482,"numberedTitle":483,"description":484,"path":485,"basePath":486,"filePath":487,"pageType":32,"hidden":13,"effectiveHidden":13,"order":135,"depth":74,"number":488,"ownAudience":62,"audiences":489,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":490},"08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/05_beurteilung_von_regressionsmodellen","05_beurteilung_von_regressionsmodellen","Beurteilung von Modellen der Regression","II.3.1.5 Beurteilung von Modellen der Regression","Unterseite zu Beurteilung von Modellen der Regression","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/05_beurteilung_von_regressionsmodellen/de","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/05_beurteilung_von_regressionsmodellen","content/08_KuenstlicheIntelligenz/02_supervisedlearning/01_grundlagen_des_sl/05_beurteilung_von_regressionsmodellen/de.md","II.3.1.5",[],[],{"key":492,"slug":493,"title":494,"cleanTitle":494,"numberedTitle":495,"description":496,"path":497,"basePath":498,"filePath":499,"pageType":32,"hidden":13,"effectiveHidden":13,"order":160,"depth":74,"number":500,"ownAudience":16,"audiences":501,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":502},"08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/01-lineare-regression","01-lineare-regression","Lineare Regression","II.3.1.6 Lineare Regression","Grundlagen der linearen Regression im überwachten Lernen.","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/01-lineare-regression/de","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/01-lineare-regression","content/08_KuenstlicheIntelligenz/02_supervisedlearning/01_grundlagen_des_sl/01-lineare-regression/de.md","II.3.1.6",[16],[],{"key":504,"slug":505,"title":506,"cleanTitle":506,"numberedTitle":507,"description":508,"path":509,"basePath":510,"filePath":511,"pageType":32,"hidden":13,"effectiveHidden":13,"order":173,"depth":74,"number":512,"ownAudience":62,"audiences":513,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":514},"08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/06_diskriminative_vs_generative_modellederklassifikation","06_diskriminative_vs_generative_modellederklassifikation","Diskriminative und Generative Modelle der Klassifikation","II.3.1.7 Diskriminative und Generative Modelle der Klassifikation","Unterseite zu Diskriminative und Generative Modelle der Klassifikation","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/06_diskriminative_vs_generative_modellederklassifikation/de","/08_kuenstlicheintelligenz/02_supervisedlearning/01_grundlagen_des_sl/06_diskriminative_vs_generative_modellederklassifikation","content/08_KuenstlicheIntelligenz/02_supervisedlearning/01_grundlagen_des_sl/06_diskriminative_vs_generative_modellederklassifikation/de.md","II.3.1.7",[],[],{"key":516,"slug":517,"title":518,"cleanTitle":518,"numberedTitle":519,"description":520,"path":521,"basePath":522,"filePath":523,"pageType":12,"hidden":13,"effectiveHidden":13,"order":33,"depth":59,"number":524,"ownAudience":62,"audiences":525,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":526},"08_kuenstlicheintelligenz/02_supervisedlearning/03_knn","03_knn","KNN","II.3.2 KNN","Unterseite zu KNN","/08_kuenstlicheintelligenz/02_supervisedlearning/03_knn/de","/08_kuenstlicheintelligenz/02_supervisedlearning/03_knn","content/08_KuenstlicheIntelligenz/02_supervisedlearning/03_knn/de.md","II.3.2",[],[],{"key":528,"slug":529,"title":530,"cleanTitle":530,"numberedTitle":531,"description":532,"path":533,"basePath":534,"filePath":535,"pageType":12,"hidden":13,"effectiveHidden":13,"order":59,"depth":59,"number":536,"ownAudience":62,"audiences":537,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":538},"08_kuenstlicheintelligenz/02_supervisedlearning/02_fnn","02_fnn","FNNs","II.3.3 FNNs","Unterseite zu FNNs","/08_kuenstlicheintelligenz/02_supervisedlearning/02_fnn/de","/08_kuenstlicheintelligenz/02_supervisedlearning/02_fnn","content/08_KuenstlicheIntelligenz/02_supervisedlearning/02_fnn/de.md","II.3.3",[16],[539,551],{"key":540,"slug":541,"title":542,"cleanTitle":542,"numberedTitle":543,"description":544,"path":545,"basePath":546,"filePath":547,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":74,"number":548,"ownAudience":16,"audiences":549,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":550},"08_kuenstlicheintelligenz/02_supervisedlearning/02_fnn/aktivierungsfunktionen","aktivierungsfunktionen","Aktivierungsfunktionen","II.3.3.1 Aktivierungsfunktionen","Unterseite zu Aktivierungsfunktionen","/08_kuenstlicheintelligenz/02_supervisedlearning/02_fnn/aktivierungsfunktionen/de","/08_kuenstlicheintelligenz/02_supervisedlearning/02_fnn/aktivierungsfunktionen","content/08_KuenstlicheIntelligenz/02_supervisedlearning/02_fnn/aktivierungsfunktionen/de.md","II.3.3.1",[16],[],{"key":552,"slug":553,"title":554,"cleanTitle":554,"numberedTitle":555,"description":556,"path":557,"basePath":558,"filePath":559,"pageType":32,"hidden":13,"effectiveHidden":13,"order":-4,"depth":74,"number":560,"ownAudience":62,"audiences":561,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":562},"08_kuenstlicheintelligenz/02_supervisedlearning/02_fnn/kunstliche-neuronen","kunstliche-neuronen","Künstliche Neuronen","II.3.3.2 Künstliche Neuronen","Unterseite zu Künstliche Neuronen","/08_kuenstlicheintelligenz/02_supervisedlearning/02_fnn/kunstliche-neuronen/de","/08_kuenstlicheintelligenz/02_supervisedlearning/02_fnn/kunstliche-neuronen","content/08_KuenstlicheIntelligenz/02_supervisedlearning/02_fnn/kunstliche-neuronen/de.md","II.3.3.2",[],[],{"key":564,"slug":565,"title":566,"cleanTitle":566,"numberedTitle":567,"description":568,"path":569,"basePath":570,"filePath":571,"pageType":58,"hidden":13,"effectiveHidden":13,"order":74,"depth":59,"number":572,"ownAudience":62,"audiences":573,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":574},"08_kuenstlicheintelligenz/02_supervisedlearning/04_svms","04_svms","SVMs","II.3.4 SVMs","Unterseite zu SVMs","/08_kuenstlicheintelligenz/02_supervisedlearning/04_svms/de","/08_kuenstlicheintelligenz/02_supervisedlearning/04_svms","content/08_KuenstlicheIntelligenz/02_supervisedlearning/04_svms/de.md","II.3.4",[],[],{"key":576,"slug":577,"title":578,"cleanTitle":578,"numberedTitle":579,"description":580,"path":581,"basePath":582,"filePath":583,"pageType":12,"hidden":13,"effectiveHidden":13,"order":160,"depth":59,"number":584,"ownAudience":62,"audiences":585,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":586},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung","05_verfahren_zur_modellanpassung","Verfahren zur Modellanpassung","II.3.5 Verfahren zur Modellanpassung","Wie Modelle ihre Parameter an Trainingsdaten anpassen","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/de.md","II.3.5",[],[587,671,743],{"key":588,"slug":589,"title":590,"cleanTitle":590,"numberedTitle":591,"description":592,"path":593,"basePath":594,"filePath":595,"pageType":58,"hidden":13,"effectiveHidden":13,"order":14,"depth":74,"number":596,"ownAudience":62,"audiences":597,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":598},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung","01_gradientenbasierte_optimierung","Gradientenbasierte Optimierung beim Training von FNNs","II.3.5.1 Gradientenbasierte Optimierung beim Training von FNNs","Parameter schrittweise mit Hilfe von Gradienten verbessern","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/de.md","II.3.5.1",[],[599,611,623,635,647,659],{"key":600,"slug":601,"title":602,"cleanTitle":602,"numberedTitle":603,"description":604,"path":605,"basePath":606,"filePath":607,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":135,"number":608,"ownAudience":62,"audiences":609,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":610},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/funktionenbeimfnn-training","funktionenbeimfnn-training","Optimierungsproblem: Training von FNNs","II.3.5.1.1 Optimierungsproblem: Training von FNNs","Wie lassen sich die Parameter eines FNNs zur Vorhersage eines Trainingsdatensatzes anpassen?","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/funktionenbeimfnn-training/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/funktionenbeimfnn-training","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/funktionenbeimfnn-training/de.md","II.3.5.1.1",[],[],{"key":612,"slug":613,"title":614,"cleanTitle":614,"numberedTitle":615,"description":616,"path":617,"basePath":618,"filePath":619,"pageType":32,"hidden":13,"effectiveHidden":13,"order":33,"depth":135,"number":620,"ownAudience":62,"audiences":621,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":622},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/01_gradientenabstieg","01_gradientenabstieg","Gradientenabstiegsverfahren","II.3.5.1.2 Gradientenabstiegsverfahren","Ein grundlegendes Verfahren zum Minimieren einer Verlustfunktion","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/01_gradientenabstieg/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/01_gradientenabstieg","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/01_gradientenabstieg/de.md","II.3.5.1.2",[],[],{"key":624,"slug":625,"title":626,"cleanTitle":626,"numberedTitle":627,"description":628,"path":629,"basePath":630,"filePath":631,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":135,"number":632,"ownAudience":62,"audiences":633,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":634},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/02_stochastischer_gradientenabstieg","02_stochastischer_gradientenabstieg","Verfahrensüberblick und Alternativen","II.3.5.1.3 Verfahrensüberblick und Alternativen","Gegenüberstellung, Verfahren höherer Ordnung","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/02_stochastischer_gradientenabstieg/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/02_stochastischer_gradientenabstieg","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/02_stochastischer_gradientenabstieg/de.md","II.3.5.1.3",[],[],{"key":636,"slug":637,"title":638,"cleanTitle":638,"numberedTitle":639,"description":640,"path":641,"basePath":642,"filePath":643,"pageType":32,"hidden":13,"effectiveHidden":13,"order":74,"depth":135,"number":644,"ownAudience":62,"audiences":645,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":646},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/03_mini_batch_verfahren","03_mini_batch_verfahren","Mini-Batch-Verfahren und Optimizer","II.3.5.1.4 Mini-Batch-Verfahren und Optimizer","Gradienten aus kleinen Gruppen von Trainingsbeispielen berechnen, Verschiedene Verwendungsansätze des Gradienten","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/03_mini_batch_verfahren/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/03_mini_batch_verfahren","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/03_mini_batch_verfahren/de.md","II.3.5.1.4",[],[],{"key":648,"slug":649,"title":650,"cleanTitle":650,"numberedTitle":651,"description":652,"path":653,"basePath":654,"filePath":655,"pageType":32,"hidden":13,"effectiveHidden":13,"order":135,"depth":135,"number":656,"ownAudience":62,"audiences":657,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":658},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/04_backpropagation","04_backpropagation","Backpropagation","II.3.5.1.5 Backpropagation","Gradienten in neuronalen Netzen effizient berechnen","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/04_backpropagation/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/04_backpropagation","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/04_backpropagation/de.md","II.3.5.1.5",[],[],{"key":660,"slug":661,"title":662,"cleanTitle":662,"numberedTitle":663,"description":664,"path":665,"basePath":666,"filePath":667,"pageType":32,"hidden":13,"effectiveHidden":13,"order":-4,"depth":135,"number":668,"ownAudience":62,"audiences":669,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":670},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/mathematischeranhang_diffrechnung","mathematischeranhang_diffrechnung","Mathematischer Anhang: Differentialrechnung","II.3.5.1.6 Mathematischer Anhang: Differentialrechnung","Unterseite zu Mathematischer Anhang: Differentialrechnung","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/mathematischeranhang_diffrechnung/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/mathematischeranhang_diffrechnung","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/01_gradientenbasierte_optimierung/mathematischeranhang_diffrechnung/de.md","II.3.5.1.6",[],[],{"key":672,"slug":673,"title":674,"cleanTitle":674,"numberedTitle":675,"description":676,"path":677,"basePath":678,"filePath":679,"pageType":58,"hidden":13,"effectiveHidden":13,"order":33,"depth":74,"number":680,"ownAudience":62,"audiences":681,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":682},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung","02_lineare_ausgleichsrechnung","Lineare Ausgleichsrechnung","II.3.5.2 Lineare Ausgleichsrechnung","Lineare Modelle so anpassen, dass Fehlerquadrate klein werden","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/de.md","II.3.5.2",[],[683,695,707,719,731],{"key":684,"slug":685,"title":686,"cleanTitle":686,"numberedTitle":687,"description":688,"path":689,"basePath":690,"filePath":691,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":135,"number":692,"ownAudience":62,"audiences":693,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":694},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/01_methode_der_kleinsten_quadrate","01_methode_der_kleinsten_quadrate","Methode der kleinsten Quadrate","II.3.5.2.1 Methode der kleinsten Quadrate","Eine Gerade oder Funktion möglichst passend an Daten legen","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/01_methode_der_kleinsten_quadrate/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/01_methode_der_kleinsten_quadrate","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/01_methode_der_kleinsten_quadrate/de.md","II.3.5.2.1",[],[],{"key":696,"slug":697,"title":698,"cleanTitle":698,"numberedTitle":699,"description":700,"path":701,"basePath":702,"filePath":703,"pageType":32,"hidden":13,"effectiveHidden":13,"order":33,"depth":135,"number":704,"ownAudience":62,"audiences":705,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":706},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/02_normalengleichungen","02_normalengleichungen","Normalengleichungen","II.3.5.2.2 Normalengleichungen","Least-Squares-Probleme als lineares Gleichungssystem lösen","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/02_normalengleichungen/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/02_normalengleichungen","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/02_normalengleichungen/de.md","II.3.5.2.2",[],[],{"key":708,"slug":709,"title":710,"cleanTitle":710,"numberedTitle":711,"description":712,"path":713,"basePath":714,"filePath":715,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":135,"number":716,"ownAudience":62,"audiences":717,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":718},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/03_qr_zerlegung","03_qr_zerlegung","QR-Zerlegung","II.3.5.2.3 QR-Zerlegung","Least-Squares-Probleme stabil mit orthogonalen Matrizen lösen","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/03_qr_zerlegung/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/03_qr_zerlegung","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/03_qr_zerlegung/de.md","II.3.5.2.3",[],[],{"key":720,"slug":721,"title":722,"cleanTitle":722,"numberedTitle":723,"description":724,"path":725,"basePath":726,"filePath":727,"pageType":32,"hidden":13,"effectiveHidden":13,"order":74,"depth":135,"number":728,"ownAudience":62,"audiences":729,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":730},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/04_householder_transformationen","04_householder_transformationen","Householder-Transformationen","II.3.5.2.4 Householder-Transformationen","Spiegelungen als Bausteine stabiler Matrixzerlegungen","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/04_householder_transformationen/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/04_householder_transformationen","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/04_householder_transformationen/de.md","II.3.5.2.4",[],[],{"key":732,"slug":733,"title":734,"cleanTitle":734,"numberedTitle":735,"description":736,"path":737,"basePath":738,"filePath":739,"pageType":32,"hidden":13,"effectiveHidden":13,"order":135,"depth":135,"number":740,"ownAudience":62,"audiences":741,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":742},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/05_regularisierung_ridge_lasso","05_regularisierung_ridge_lasso","Regularisierung: Ridge und Lasso","II.3.5.2.5 Regularisierung: Ridge und Lasso","Wie Ridge- und Lasso-Regression Least Squares erweitern","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/05_regularisierung_ridge_lasso/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/05_regularisierung_ridge_lasso","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/02_lineare_ausgleichsrechnung/05_regularisierung_ridge_lasso/de.md","II.3.5.2.5",[],[],{"key":744,"slug":745,"title":746,"cleanTitle":746,"numberedTitle":747,"description":748,"path":749,"basePath":750,"filePath":751,"pageType":58,"hidden":13,"effectiveHidden":13,"order":59,"depth":74,"number":752,"ownAudience":62,"audiences":753,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":754},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung","03_statistische_parameterschaetzung","Statistische Parameterschätzung","II.3.5.3 Statistische Parameterschätzung","Parameter als Erklärung für beobachtete Daten bestimmen","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/de.md","II.3.5.3",[],[755,767,779,791,803],{"key":756,"slug":757,"title":758,"cleanTitle":758,"numberedTitle":759,"description":760,"path":761,"basePath":762,"filePath":763,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":135,"number":764,"ownAudience":62,"audiences":765,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":766},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/01_maximum_likelihood_schaetzung","01_maximum_likelihood_schaetzung","Maximum-Likelihood-Schätzung","II.3.5.3.1 Maximum-Likelihood-Schätzung","Parameter wählen, unter denen die beobachteten Daten möglichst wahrscheinlich sind","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/01_maximum_likelihood_schaetzung/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/01_maximum_likelihood_schaetzung","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/01_maximum_likelihood_schaetzung/de.md","II.3.5.3.1",[],[],{"key":768,"slug":769,"title":770,"cleanTitle":770,"numberedTitle":771,"description":772,"path":773,"basePath":774,"filePath":775,"pageType":32,"hidden":13,"effectiveHidden":13,"order":33,"depth":135,"number":776,"ownAudience":62,"audiences":777,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":778},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/02_latente_variablen","02_latente_variablen","Latente Variablen","II.3.5.3.2 Latente Variablen","Verborgene Größen in statistischen Modellen berücksichtigen","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/02_latente_variablen/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/02_latente_variablen","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/02_latente_variablen/de.md","II.3.5.3.2",[],[],{"key":780,"slug":781,"title":782,"cleanTitle":782,"numberedTitle":783,"description":784,"path":785,"basePath":786,"filePath":787,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":135,"number":788,"ownAudience":62,"audiences":789,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":790},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/03_e_schritt","03_e_schritt","E-Schritt","II.3.5.3.3 E-Schritt","Erwartungen über verborgene Variablen berechnen","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/03_e_schritt/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/03_e_schritt","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/03_e_schritt/de.md","II.3.5.3.3",[],[],{"key":792,"slug":793,"title":794,"cleanTitle":794,"numberedTitle":795,"description":796,"path":797,"basePath":798,"filePath":799,"pageType":32,"hidden":13,"effectiveHidden":13,"order":74,"depth":135,"number":800,"ownAudience":62,"audiences":801,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":802},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/04_m_schritt","04_m_schritt","M-Schritt","II.3.5.3.4 M-Schritt","Parameter anhand der erwarteten verborgenen Informationen verbessern","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/04_m_schritt/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/04_m_schritt","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/04_m_schritt/de.md","II.3.5.3.4",[],[],{"key":804,"slug":805,"title":806,"cleanTitle":806,"numberedTitle":807,"description":808,"path":809,"basePath":810,"filePath":811,"pageType":32,"hidden":13,"effectiveHidden":13,"order":135,"depth":135,"number":812,"ownAudience":62,"audiences":813,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":814},"08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/05_em_algorithmus","05_em_algorithmus","EM-Algorithmus","II.3.5.3.5 EM-Algorithmus","Abwechselnd verborgene Informationen und Parameter schätzen","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/05_em_algorithmus/de","/08_kuenstlicheintelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/05_em_algorithmus","content/08_KuenstlicheIntelligenz/02_supervisedlearning/05_verfahren_zur_modellanpassung/03_statistische_parameterschaetzung/05_em_algorithmus/de.md","II.3.5.3.5",[],[],{"key":816,"slug":817,"title":818,"cleanTitle":818,"numberedTitle":819,"description":820,"path":821,"basePath":822,"filePath":823,"pageType":32,"hidden":13,"effectiveHidden":13,"order":173,"depth":59,"number":824,"ownAudience":62,"audiences":825,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":826},"08_kuenstlicheintelligenz/02_supervisedlearning/modelzoo","modelzoo","Modell Zoo","II.3.6 Modell Zoo","Unterseite zu Modell Zoo","/08_kuenstlicheintelligenz/02_supervisedlearning/modelzoo/de","/08_kuenstlicheintelligenz/02_supervisedlearning/modelzoo","content/08_KuenstlicheIntelligenz/02_supervisedlearning/modelzoo/de.md","II.3.6",[],[],{"key":828,"slug":829,"title":830,"cleanTitle":830,"numberedTitle":831,"description":832,"path":833,"basePath":834,"filePath":835,"pageType":12,"hidden":13,"effectiveHidden":13,"order":74,"depth":33,"number":836,"ownAudience":62,"audiences":837,"coverDeLight":284,"coverEnLight":285,"coverDeDark":286,"coverEnDark":287,"children":838},"08_kuenstlicheintelligenz/05_unsupervisedlearning","05_unsupervisedlearning","Unüberwachtes Lernen","II.4 Unüberwachtes Lernen","Unterseite zu Unüberwachtem Lernen","/08_kuenstlicheintelligenz/05_unsupervisedlearning/de","/08_kuenstlicheintelligenz/05_unsupervisedlearning","content/08_KuenstlicheIntelligenz/05_unsupervisedlearning/de.md","II.4",[16],[839,851,863,875,937],{"key":840,"slug":841,"title":842,"cleanTitle":842,"numberedTitle":843,"description":844,"path":845,"basePath":846,"filePath":847,"pageType":58,"hidden":13,"effectiveHidden":13,"order":-4,"depth":59,"number":848,"ownAudience":62,"audiences":849,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":850},"08_kuenstlicheintelligenz/05_unsupervisedlearning/anomalydetection","anomalydetection","Anomalieerkennung","II.4.1 Anomalieerkennung","Unterseite zu Anomalieerkennung","/08_kuenstlicheintelligenz/05_unsupervisedlearning/anomalydetection/de","/08_kuenstlicheintelligenz/05_unsupervisedlearning/anomalydetection","content/08_KuenstlicheIntelligenz/05_unsupervisedlearning/anomalydetection/de.md","II.4.1",[],[],{"key":852,"slug":853,"title":854,"cleanTitle":854,"numberedTitle":855,"description":856,"path":857,"basePath":858,"filePath":859,"pageType":58,"hidden":13,"effectiveHidden":13,"order":-4,"depth":59,"number":860,"ownAudience":62,"audiences":861,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":862},"08_kuenstlicheintelligenz/05_unsupervisedlearning/associationrulelearning","associationrulelearning","Assoziationsregel-Lernen","II.4.2 Assoziationsregel-Lernen","Unterseite zu Assoziationsregel-Lernen","/08_kuenstlicheintelligenz/05_unsupervisedlearning/associationrulelearning/de","/08_kuenstlicheintelligenz/05_unsupervisedlearning/associationrulelearning","content/08_KuenstlicheIntelligenz/05_unsupervisedlearning/associationrulelearning/de.md","II.4.2",[],[],{"key":864,"slug":865,"title":866,"cleanTitle":866,"numberedTitle":867,"description":868,"path":869,"basePath":870,"filePath":871,"pageType":32,"hidden":13,"effectiveHidden":13,"order":-4,"depth":59,"number":872,"ownAudience":62,"audiences":873,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":874},"08_kuenstlicheintelligenz/05_unsupervisedlearning/aufgabendesul","aufgabendesul","Aufgaben im Unsupervised Learning","II.4.3 Aufgaben im Unsupervised Learning","Unterseite zu Aufgaben im Unsupervised Learning","/08_kuenstlicheintelligenz/05_unsupervisedlearning/aufgabendesul/de","/08_kuenstlicheintelligenz/05_unsupervisedlearning/aufgabendesul","content/08_KuenstlicheIntelligenz/05_unsupervisedlearning/aufgabendesul/de.md","II.4.3",[],[],{"key":876,"slug":877,"title":878,"cleanTitle":878,"numberedTitle":879,"description":880,"path":881,"basePath":882,"filePath":883,"pageType":58,"hidden":13,"effectiveHidden":13,"order":-4,"depth":59,"number":884,"ownAudience":62,"audiences":885,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":886},"08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering","clustering","Clustering","II.4.4 Clustering","Unterseite zu Clustering","/08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/de","/08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering","content/08_KuenstlicheIntelligenz/05_unsupervisedlearning/clustering/de.md","II.4.4",[16],[887,899,911,924],{"key":888,"slug":889,"title":890,"cleanTitle":890,"numberedTitle":891,"description":892,"path":893,"basePath":894,"filePath":895,"pageType":32,"hidden":13,"effectiveHidden":13,"order":14,"depth":74,"number":896,"ownAudience":62,"audiences":897,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":898},"08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/dbscan","dbscan","DBSCAN","II.4.4.1 DBSCAN","Unterseite zu DBSCAN","/08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/dbscan/de","/08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/dbscan","content/08_KuenstlicheIntelligenz/05_unsupervisedlearning/clustering/dbscan/de.md","II.4.4.1",[],[],{"key":900,"slug":901,"title":902,"cleanTitle":902,"numberedTitle":903,"description":904,"path":905,"basePath":906,"filePath":907,"pageType":32,"hidden":13,"effectiveHidden":13,"order":33,"depth":74,"number":908,"ownAudience":62,"audiences":909,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":910},"08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/kmeans","kmeans","Das k-Means Verfahren","II.4.4.2 Das k-Means Verfahren","Unterseite zu Das k-Means Verfahren","/08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/kmeans/de","/08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/kmeans","content/08_KuenstlicheIntelligenz/05_unsupervisedlearning/clustering/kmeans/de.md","II.4.4.2",[],[],{"key":912,"slug":913,"title":914,"cleanTitle":915,"numberedTitle":916,"description":917,"path":918,"basePath":919,"filePath":920,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":74,"number":921,"ownAudience":16,"audiences":922,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":923},"08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/01-einfuehrung-kmeans","01-einfuehrung-kmeans","01-Einführung in K-Means Clustering","Einführung in K-Means Clustering","II.4.4.3 Einführung in K-Means Clustering","Grundlagen und interaktive Visualisierung des K-Means Algorithmus.","/08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/01-einfuehrung-kmeans/de","/08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/01-einfuehrung-kmeans","content/08_KuenstlicheIntelligenz/05_unsupervisedlearning/clustering/01-einfuehrung-kmeans/de.md","II.4.4.3",[16],[],{"key":925,"slug":926,"title":927,"cleanTitle":928,"numberedTitle":929,"description":930,"path":931,"basePath":932,"filePath":933,"pageType":32,"hidden":13,"effectiveHidden":13,"order":74,"depth":74,"number":934,"ownAudience":16,"audiences":935,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"children":936},"08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/02-einfuehrung-dbscan","02-einfuehrung-dbscan","02-Einführung in DBSCAN Clustering","Einführung in DBSCAN Clustering","II.4.4.4 Einführung in DBSCAN Clustering","Grundlagen und interaktive Visualisierung des DBSCAN Algorithmus.","/08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/02-einfuehrung-dbscan/de","/08_kuenstlicheintelligenz/05_unsupervisedlearning/clustering/02-einfuehrung-dbscan","content/08_KuenstlicheIntelligenz/05_unsupervisedlearning/clustering/02-einfuehrung-dbscan/de.md","II.4.4.4",[16],[],{"key":938,"slug":939,"title":940,"cleanTitle":940,"numberedTitle":941,"description":942,"path":943,"basePath":944,"filePath":945,"pageType":58,"hidden":13,"effectiveHidden":13,"order":-4,"depth":59,"number":946,"ownAudience":62,"audiences":947,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":948},"08_kuenstlicheintelligenz/05_unsupervisedlearning/dimensionalityreduction","dimensionalityreduction","Dimensionsreduktion","II.4.5 Dimensionsreduktion","Unterseite zu Dimensionsreduktion","/08_kuenstlicheintelligenz/05_unsupervisedlearning/dimensionalityreduction/de","/08_kuenstlicheintelligenz/05_unsupervisedlearning/dimensionalityreduction","content/08_KuenstlicheIntelligenz/05_unsupervisedlearning/dimensionalityreduction/de.md","II.4.5",[],[],{"key":950,"slug":951,"title":952,"cleanTitle":952,"numberedTitle":953,"description":954,"path":955,"basePath":956,"filePath":957,"pageType":12,"hidden":13,"effectiveHidden":13,"order":135,"depth":33,"number":958,"ownAudience":62,"audiences":959,"coverDeLight":284,"coverEnLight":285,"coverDeDark":286,"coverEnDark":287,"children":960},"08_kuenstlicheintelligenz/06_reinforcementlearning","06_reinforcementlearning","Verstärkendes Lernen","II.5 Verstärkendes Lernen","Unterseite zu Verstärkendes Lernen","/08_kuenstlicheintelligenz/06_reinforcementlearning/de","/08_kuenstlicheintelligenz/06_reinforcementlearning","content/08_KuenstlicheIntelligenz/06_reinforcementlearning/de.md","II.5",[16],[961,973,986],{"key":962,"slug":963,"title":964,"cleanTitle":964,"numberedTitle":965,"description":966,"path":967,"basePath":968,"filePath":969,"pageType":58,"hidden":13,"effectiveHidden":13,"order":14,"depth":59,"number":970,"ownAudience":62,"audiences":971,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":972},"08_kuenstlicheintelligenz/06_reinforcementlearning/rlfoundations","rlfoundations","Grundlagen des Reinforcement Learnings","II.5.1 Grundlagen des Reinforcement Learnings","Unterseite zu Grundlagen des Reinforcement Learnings","/08_kuenstlicheintelligenz/06_reinforcementlearning/rlfoundations/de","/08_kuenstlicheintelligenz/06_reinforcementlearning/rlfoundations","content/08_KuenstlicheIntelligenz/06_reinforcementlearning/rlfoundations/de.md","II.5.1",[],[],{"key":974,"slug":975,"title":976,"cleanTitle":977,"numberedTitle":978,"description":979,"path":980,"basePath":981,"filePath":982,"pageType":58,"hidden":13,"effectiveHidden":13,"order":33,"depth":59,"number":983,"ownAudience":16,"audiences":984,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":985},"08_kuenstlicheintelligenz/06_reinforcementlearning/02-werte-lernen","02-werte-lernen","02-Einführung in value-based Reinforcement Learning","Einführung in value-based Reinforcement Learning","II.5.2 Einführung in value-based Reinforcement Learning","Grundlagen und interaktive Visualisierung von Q-Learning für das NIM-Spiel.","/08_kuenstlicheintelligenz/06_reinforcementlearning/02-werte-lernen/de","/08_kuenstlicheintelligenz/06_reinforcementlearning/02-werte-lernen","content/08_KuenstlicheIntelligenz/06_reinforcementlearning/02-werte-lernen/de.md","II.5.2",[16],[],{"key":987,"slug":988,"title":989,"cleanTitle":990,"numberedTitle":991,"description":992,"path":993,"basePath":994,"filePath":995,"pageType":32,"hidden":13,"effectiveHidden":13,"order":59,"depth":59,"number":996,"ownAudience":16,"audiences":997,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":998},"08_kuenstlicheintelligenz/06_reinforcementlearning/01-strategien-lernen","01-strategien-lernen","01-Einführung in policy-based Reinforcement Learning","Einführung in policy-based Reinforcement Learning","II.5.3 Einführung in policy-based Reinforcement Learning","Grundlagen und interaktive Visualisierung einer Adaption von MENACE für das NIM-Spiel.","/08_kuenstlicheintelligenz/06_reinforcementlearning/01-strategien-lernen/de","/08_kuenstlicheintelligenz/06_reinforcementlearning/01-strategien-lernen","content/08_KuenstlicheIntelligenz/06_reinforcementlearning/01-strategien-lernen/de.md","II.5.3",[16],[],{"key":1000,"slug":1000,"title":1001,"cleanTitle":1001,"numberedTitle":1001,"description":62,"path":62,"basePath":1002,"filePath":62,"pageType":62,"hidden":13,"effectiveHidden":13,"order":-4,"depth":14,"number":62,"ownAudience":62,"audiences":1003,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":1004},"_info","Info","/_info",[16],[1005,1016,1027,1038],{"key":1006,"slug":1007,"title":1008,"cleanTitle":1008,"numberedTitle":1008,"description":1009,"path":1010,"basePath":1011,"filePath":1012,"pageType":32,"hidden":13,"effectiveHidden":13,"order":1013,"depth":33,"number":62,"ownAudience":16,"audiences":1014,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":1015},"_info/10_datenschutz","10_datenschutz","Datenschutzerklärung","Informationen zum Datenschutz und zur Verarbeitung personenbezogener Daten.","/_info/10_datenschutz/de","/_info/10_datenschutz","content/_Info/10_Datenschutz/de.md",10,[16],[],{"key":1017,"slug":1018,"title":1019,"cleanTitle":1019,"numberedTitle":1019,"description":1020,"path":1021,"basePath":1022,"filePath":1023,"pageType":32,"hidden":13,"effectiveHidden":13,"order":1024,"depth":33,"number":62,"ownAudience":16,"audiences":1025,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":1026},"_info/11_impressum","11_impressum","Impressum","Impressum fuer das interaktive Buch.","/_info/11_impressum/de","/_info/11_impressum","content/_Info/11_Impressum/de.md",11,[16],[],{"key":1028,"slug":1029,"title":1030,"cleanTitle":1030,"numberedTitle":1030,"description":1031,"path":1032,"basePath":1033,"filePath":1034,"pageType":32,"hidden":13,"effectiveHidden":13,"order":1035,"depth":33,"number":62,"ownAudience":16,"audiences":1036,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":1037},"_info/12_lizenzenunddrittanbieter","12_lizenzenunddrittanbieter","Lizenzen & Drittanbieter","Übersicht der verwendeten Open-Source-Komponenten, Lizenzen und Datenschutz-Hinweise.","/_info/12_lizenzenunddrittanbieter/de","/_info/12_lizenzenunddrittanbieter","content/_Info/12_LizenzenUndDrittanbieter/de.md",12,[16],[],{"key":1039,"slug":1040,"title":1041,"cleanTitle":1041,"numberedTitle":1041,"description":1042,"path":1043,"basePath":1044,"filePath":1045,"pageType":58,"hidden":13,"effectiveHidden":13,"order":1046,"depth":33,"number":62,"ownAudience":16,"audiences":1047,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":1048},"_info/13_mailinglisteundnewsletter","13_mailinglisteundnewsletter","Mailingliste und Newsletter","Benachrichtigung über Neuigkeiten","/_info/13_mailinglisteundnewsletter/de","/_info/13_mailinglisteundnewsletter","content/_Info/13_MailingListeUndNewsLetter/de.md",13,[16],[1049,1061],{"key":1050,"slug":1051,"title":1052,"cleanTitle":1052,"numberedTitle":1053,"description":1054,"path":1055,"basePath":1056,"filePath":1057,"pageType":32,"hidden":13,"effectiveHidden":13,"order":-4,"depth":59,"number":1058,"ownAudience":16,"audiences":1059,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":1060},"_info/13_mailinglisteundnewsletter/ifk","ifk","Mailingliste Informatiklehrer Köln und Umgebung","1 Mailingliste Informatiklehrer Köln und Umgebung","Diskussion und Austausch in der weiteren Kölner Region","/_info/13_mailinglisteundnewsletter/ifk/de","/_info/13_mailinglisteundnewsletter/ifk","content/_Info/13_MailingListeUndNewsLetter/ifk/de.md","1",[16],[],{"key":1062,"slug":1063,"title":1064,"cleanTitle":1064,"numberedTitle":1065,"description":1066,"path":1067,"basePath":1068,"filePath":1069,"pageType":32,"hidden":13,"effectiveHidden":13,"order":-4,"depth":59,"number":1070,"ownAudience":16,"audiences":1071,"coverDeLight":62,"coverEnLight":62,"coverDeDark":62,"coverEnDark":62,"children":1072},"_info/13_mailinglisteundnewsletter/newsletter","newsletter","Newsletter","2 Newsletter","Anmeldung zum Newsletter","/_info/13_mailinglisteundnewsletter/newsletter/de","/_info/13_mailinglisteundnewsletter/newsletter","content/_Info/13_MailingListeUndNewsLetter/newsletter/de.md","2",[16],[],{"_path":329,"_dir":325,"_draft":13,"_partial":13,"_locale":62,"title":326,"description":328,"navigation":1074,"author":1075,"audience":16,"pageType":32,"order":74,"coverDeLight":18,"coverEnLight":19,"coverDeDark":20,"coverEnDark":21,"body":1076,"_type":2175,"_id":2176,"_source":32,"_file":2177,"_stem":2178,"_extension":2179},true,"HM_JDS",{"type":1077,"children":1078,"toc":2153},"root",[1079,1087,1126,1131,1136,1148,1164,1190,1197,1202,1207,1212,1221,1227,1232,1297,1302,1307,1313,1318,1365,1377,1383,1393,1398,1403,1408,1425,1431,1436,1507,1512,1529,1535,1540,1545,1564,1570,1575,1582,1587,1605,1610,1616,1635,1640,1645,1687,1693,1698,1703,1708,1751,1756,1762,1767,1773,1807,1812,1937,1943,1951,1957,1962,1967,2019,2024,2030,2035,2040,2073,2078,2083,2089,2094,2115,2120,2148],{"type":1080,"tag":1081,"props":1082,"children":1084},"element","h1",{"id":1083},"feature-extraction",[1085],{"type":1086,"value":326},"text",{"type":1080,"tag":1088,"props":1089,"children":1090},"p",{},[1091,1093,1099,1101,1108,1110,1116,1118,1124],{"type":1086,"value":1092},"Auf der vorherigen Seite wurden bereits ",{"type":1080,"tag":1094,"props":1095,"children":1096},"strong",{},[1097],{"type":1086,"value":1098},"Merkmale",{"type":1086,"value":1100}," eines Datensatzes betrachtet, zum Beispiel ",{"type":1080,"tag":1102,"props":1103,"children":1105},"code",{"className":1104},[],[1106],{"type":1086,"value":1107},"duration_s",{"type":1086,"value":1109},", ",{"type":1080,"tag":1102,"props":1111,"children":1113},{"className":1112},[],[1114],{"type":1086,"value":1115},"dominant_frequency_hz",{"type":1086,"value":1117}," oder ",{"type":1080,"tag":1102,"props":1119,"children":1121},{"className":1120},[],[1122],{"type":1086,"value":1123},"spectral_centroid_hz",{"type":1086,"value":1125},".",{"type":1080,"tag":1088,"props":1127,"children":1128},{},[1129],{"type":1086,"value":1130},"Dabei stellt sich nun eine wichtige Frage:",{"type":1080,"tag":1088,"props":1132,"children":1133},{},[1134],{"type":1086,"value":1135},"Woher kommen solche Merkmale eigentlich?",{"type":1080,"tag":1088,"props":1137,"children":1138},{},[1139,1141,1146],{"type":1086,"value":1140},"Beim maschinellen Lernen liegen Daten oft nicht sofort als fertige Tabelle mit Spalten und Zahlen vor. Häufig beginnt man mit ",{"type":1080,"tag":1094,"props":1142,"children":1143},{},[1144],{"type":1086,"value":1145},"Rohdaten",{"type":1086,"value":1147},", etwa mit Texten, Bildern, Videos, Messwertreihen oder Audioaufnahmen. Damit ein Lernverfahren damit arbeiten kann, müssen daraus oft erst geeignete Merkmale gewonnen werden.",{"type":1080,"tag":1088,"props":1149,"children":1150},{},[1151,1153,1157,1158,1163],{"type":1086,"value":1152},"Diesen Verarbeitungsschritt nennt man ",{"type":1080,"tag":1094,"props":1154,"children":1155},{},[1156],{"type":1086,"value":326},{"type":1086,"value":1117},{"type":1080,"tag":1094,"props":1159,"children":1160},{},[1161],{"type":1086,"value":1162},"Merkmalsextraktion",{"type":1086,"value":1125},{"type":1080,"tag":1165,"props":1166,"children":1167},"blockquote",{},[1168],{"type":1080,"tag":1088,"props":1169,"children":1170},{},[1171,1175,1177,1182,1184,1189],{"type":1080,"tag":1094,"props":1172,"children":1173},{},[1174],{"type":1086,"value":326},{"type":1086,"value":1176}," ist der Schritt, in dem aus Rohdaten beschreibende Merkmale gewonnen werden.\nDiese Merkmale heißen im maschinellen Lernen ",{"type":1080,"tag":1094,"props":1178,"children":1179},{},[1180],{"type":1086,"value":1181},"Features",{"type":1086,"value":1183}," und erscheinen in einer Tabelle als ",{"type":1080,"tag":1094,"props":1185,"children":1186},{},[1187],{"type":1086,"value":1188},"statistische Variablen",{"type":1086,"value":1125},{"type":1080,"tag":1191,"props":1192,"children":1194},"h2",{"id":1193},"rohdaten-und-merkmale",[1195],{"type":1086,"value":1196},"Rohdaten und Merkmale",{"type":1080,"tag":1088,"props":1198,"children":1199},{},[1200],{"type":1086,"value":1201},"Rohdaten können sehr unterschiedlich aussehen. Ein Text besteht zunächst aus Zeichen und Wörtern, ein Bild aus Pixeln, ein Video aus vielen aufeinanderfolgenden Bildern und eine Audioaufnahme aus einem zeitlich veränderlichen Signal.",{"type":1080,"tag":1088,"props":1203,"children":1204},{},[1205],{"type":1086,"value":1206},"Solche Rohdaten sind zwar oft bereits digital gespeichert, aber sie liegen noch nicht unbedingt in einer Form vor, in der ein Lernverfahren direkt sinnvoll damit arbeiten kann.",{"type":1080,"tag":1088,"props":1208,"children":1209},{},[1210],{"type":1086,"value":1211},"Erst durch geeignete Verarbeitungsschritte werden daraus Merkmale, mit denen man später Datenpunkte vergleichen, statistische Auswertungen vornehmen oder ein Modell trainieren kann.",{"type":1080,"tag":1213,"props":1214,"children":1215},"key-point-box",{},[1216],{"type":1080,"tag":1088,"props":1217,"children":1218},{},[1219],{"type":1086,"value":1220},"Rohdaten sind der Ausgangspunkt.\nMerkmale sind die beschreibenden Größen, die aus den Rohdaten gewonnen werden.",{"type":1080,"tag":1191,"props":1222,"children":1224},{"id":1223},"welche-arten-von-rohdaten-kommen-häufig-vor",[1225],{"type":1086,"value":1226},"Welche Arten von Rohdaten kommen häufig vor?",{"type":1080,"tag":1088,"props":1228,"children":1229},{},[1230],{"type":1086,"value":1231},"In der Informatik und besonders in der Data Science begegnet man vielen unterschiedlichen Arten von Rohdaten, zum Beispiel:",{"type":1080,"tag":1233,"props":1234,"children":1235},"ul",{},[1236,1247,1257,1267,1277,1287],{"type":1080,"tag":1237,"props":1238,"children":1239},"li",{},[1240,1245],{"type":1080,"tag":1094,"props":1241,"children":1242},{},[1243],{"type":1086,"value":1244},"Textdaten",{"type":1086,"value":1246},", etwa Nachrichtenartikel, E-Mails oder Chatnachrichten",{"type":1080,"tag":1237,"props":1248,"children":1249},{},[1250,1255],{"type":1080,"tag":1094,"props":1251,"children":1252},{},[1253],{"type":1086,"value":1254},"Bilddaten",{"type":1086,"value":1256},", etwa Fotos, Satellitenbilder oder medizinische Aufnahmen",{"type":1080,"tag":1237,"props":1258,"children":1259},{},[1260,1265],{"type":1080,"tag":1094,"props":1261,"children":1262},{},[1263],{"type":1086,"value":1264},"Audiodaten",{"type":1086,"value":1266},", etwa Sprache, Musik oder Tierstimmen",{"type":1080,"tag":1237,"props":1268,"children":1269},{},[1270,1275],{"type":1080,"tag":1094,"props":1271,"children":1272},{},[1273],{"type":1086,"value":1274},"Videodaten",{"type":1086,"value":1276},", also Folgen von Bildern mit Ton",{"type":1080,"tag":1237,"props":1278,"children":1279},{},[1280,1285],{"type":1080,"tag":1094,"props":1281,"children":1282},{},[1283],{"type":1086,"value":1284},"Zeitreihendaten",{"type":1086,"value":1286},", etwa Wetterwerte, Aktienkurse oder Sensorsignale",{"type":1080,"tag":1237,"props":1288,"children":1289},{},[1290,1295],{"type":1080,"tag":1094,"props":1291,"children":1292},{},[1293],{"type":1086,"value":1294},"tabellarische Daten",{"type":1086,"value":1296},", etwa Messdaten, Umfragedaten oder Verkaufsdaten",{"type":1080,"tag":1088,"props":1298,"children":1299},{},[1300],{"type":1086,"value":1301},"Manche Daten liegen bereits in Tabellenform vor. Andere müssen erst so verarbeitet werden, dass daraus geeignete Merkmale entstehen.",{"type":1080,"tag":1088,"props":1303,"children":1304},{},[1305],{"type":1086,"value":1306},"Bei tabellarischen Daten können manche Merkmale schon direkt als Spalten vorliegen. Bei Audio, Bildern oder Texten ist das oft nicht so. Dort braucht man meist zusätzliche Verarbeitungsschritte, um aus den Rohdaten beschreibende Merkmale zu gewinnen.",{"type":1080,"tag":1191,"props":1308,"children":1310},{"id":1309},"von-rohdaten-zu-merkmalen",[1311],{"type":1086,"value":1312},"Von Rohdaten zu Merkmalen",{"type":1080,"tag":1088,"props":1314,"children":1315},{},[1316],{"type":1086,"value":1317},"Wie Feature Extraction genau aussieht, hängt von der Art der Daten ab.",{"type":1080,"tag":1233,"props":1319,"children":1320},{},[1321,1333,1344,1355],{"type":1080,"tag":1237,"props":1322,"children":1323},{},[1324,1326,1331],{"type":1086,"value":1325},"Bei ",{"type":1080,"tag":1094,"props":1327,"children":1328},{},[1329],{"type":1086,"value":1330},"Texten",{"type":1086,"value":1332}," kann man zum Beispiel Wortanzahlen, Worthäufigkeiten oder Satzlängen bestimmen.",{"type":1080,"tag":1237,"props":1334,"children":1335},{},[1336,1337,1342],{"type":1086,"value":1325},{"type":1080,"tag":1094,"props":1338,"children":1339},{},[1340],{"type":1086,"value":1341},"Bildern",{"type":1086,"value":1343}," kann man zum Beispiel Farben, Kanten oder Formen beschreiben.",{"type":1080,"tag":1237,"props":1345,"children":1346},{},[1347,1348,1353],{"type":1086,"value":1325},{"type":1080,"tag":1094,"props":1349,"children":1350},{},[1351],{"type":1086,"value":1352},"Zeitreihen",{"type":1086,"value":1354}," kann man zum Beispiel Mittelwerte, Trends oder Schwankungen betrachten.",{"type":1080,"tag":1237,"props":1356,"children":1357},{},[1358,1359,1363],{"type":1086,"value":1325},{"type":1080,"tag":1094,"props":1360,"children":1361},{},[1362],{"type":1086,"value":1264},{"type":1086,"value":1364}," kann man zum Beispiel Dauer, Energie oder typische Frequenzbereiche messen.",{"type":1080,"tag":1088,"props":1366,"children":1367},{},[1368,1370,1375],{"type":1086,"value":1369},"In diesem Kapitel geht es um ",{"type":1080,"tag":1094,"props":1371,"children":1372},{},[1373],{"type":1086,"value":1374},"Audioaufnahmen von Froschrufen",{"type":1086,"value":1376},". An ihnen lässt sich gut erkennen, wie aus einer Rohaufnahme Schritt für Schritt Merkmale für einen Datensatz entstehen.",{"type":1080,"tag":1191,"props":1378,"children":1380},{"id":1379},"beispiel-froschrufe-als-audiodaten",[1381],{"type":1086,"value":1382},"Beispiel: Froschrufe als Audiodaten",{"type":1080,"tag":1088,"props":1384,"children":1385},{},[1386,1388,1392],{"type":1086,"value":1387},"In unserem Beispiel liegen die Rohdaten nicht sofort als fertige Tabellenzeile vor, sondern als ",{"type":1080,"tag":1094,"props":1389,"children":1390},{},[1391],{"type":1086,"value":1374},{"type":1086,"value":1125},{"type":1080,"tag":1088,"props":1394,"children":1395},{},[1396],{"type":1086,"value":1397},"Eine solche Aufnahme ist zunächst ein digitales Signal. Es beschreibt, wie sich der Schalldruck im Zeitverlauf ändert. Im Computer wird dieses Signal als Folge vieler Abtastwerte gespeichert.",{"type":1080,"tag":1088,"props":1399,"children":1400},{},[1401],{"type":1086,"value":1402},"Diese vielen einzelnen Werte sind zwar bereits Zahlen, aber noch nicht automatisch die Merkmale, mit denen man verschiedene Froschrufe gut vergleichen kann. Deshalb werden aus der Aufnahme weitere beschreibende Größen berechnet.",{"type":1080,"tag":1088,"props":1404,"children":1405},{},[1406],{"type":1086,"value":1407},"Genau darin besteht hier die Feature Extraction: Aus dem Rohsignal werden Merkmale gewonnen, die wichtige Eigenschaften des Rufs zusammenfassen.",{"type":1080,"tag":1213,"props":1409,"children":1412},{"title":1410,"label":1411},"Was ist hier ein Merkmal?","Merke",[1413],{"type":1080,"tag":1088,"props":1414,"children":1415},{},[1416,1418,1423],{"type":1086,"value":1417},"Ein ",{"type":1080,"tag":1094,"props":1419,"children":1420},{},[1421],{"type":1086,"value":1422},"Merkmal",{"type":1086,"value":1424}," in diesem Zusammenhang ist eine beschreibende Eigenschaft eines Audiosegments, zum Beispiel seine Dauer, seine Energie oder eine charakteristische Frequenz.",{"type":1080,"tag":1191,"props":1426,"children":1428},{"id":1427},"welche-merkmale-werden-hier-betrachtet",[1429],{"type":1086,"value":1430},"Welche Merkmale werden hier betrachtet?",{"type":1080,"tag":1088,"props":1432,"children":1433},{},[1434],{"type":1086,"value":1435},"Bei diesem Datensatz werden zum Beispiel folgende Merkmale betrachtet:",{"type":1080,"tag":1233,"props":1437,"children":1438},{},[1439,1453,1467,1480,1493],{"type":1080,"tag":1237,"props":1440,"children":1441},{},[1442,1447,1451],{"type":1080,"tag":1102,"props":1443,"children":1445},{"className":1444},[],[1446],{"type":1086,"value":1107},{"type":1080,"tag":1448,"props":1449,"children":1450},"br",{},[],{"type":1086,"value":1452},"\nDauer des Audiosegments in Sekunden",{"type":1080,"tag":1237,"props":1454,"children":1455},{},[1456,1462,1465],{"type":1080,"tag":1102,"props":1457,"children":1459},{"className":1458},[],[1460],{"type":1086,"value":1461},"rms_energy",{"type":1080,"tag":1448,"props":1463,"children":1464},{},[],{"type":1086,"value":1466},"\nEin Maß für die mittlere Signalstärke oder Energie",{"type":1080,"tag":1237,"props":1468,"children":1469},{},[1470,1475,1478],{"type":1080,"tag":1102,"props":1471,"children":1473},{"className":1472},[],[1474],{"type":1086,"value":1123},{"type":1080,"tag":1448,"props":1476,"children":1477},{},[],{"type":1086,"value":1479},"\nDer spektrale Schwerpunkt in Hertz, also vereinfacht gesagt der \"Helligkeitsschwerpunkt\" des Klangs",{"type":1080,"tag":1237,"props":1481,"children":1482},{},[1483,1488,1491],{"type":1080,"tag":1102,"props":1484,"children":1486},{"className":1485},[],[1487],{"type":1086,"value":1115},{"type":1080,"tag":1448,"props":1489,"children":1490},{},[],{"type":1086,"value":1492},"\nDie dominante Frequenz in Hertz, also die Frequenz mit besonders starker Ausprägung im Signal",{"type":1080,"tag":1237,"props":1494,"children":1495},{},[1496,1502,1505],{"type":1080,"tag":1102,"props":1497,"children":1499},{"className":1498},[],[1500],{"type":1086,"value":1501},"onset_count",{"type":1080,"tag":1448,"props":1503,"children":1504},{},[],{"type":1086,"value":1506},"\nAnzahl erkannter Einsätze oder Impulse",{"type":1080,"tag":1088,"props":1508,"children":1509},{},[1510],{"type":1086,"value":1511},"Diese Merkmale werden später als Spalten in einer Tabelle gespeichert. Jedes Audiosegment liefert also eine neue Zeile mit konkreten Merkmalswerten.",{"type":1080,"tag":1513,"props":1514,"children":1516},"definition-box",{"title":1515},"Feature-Extraction (Merkmalsextraktion)",[1517],{"type":1080,"tag":1088,"props":1518,"children":1519},{},[1520,1522,1527],{"type":1086,"value":1521},"Feature Extraction bezeichnet den Verarbeitungsschritt, bei dem aus Rohdaten beschreibende Merkmale gewonnen werden, die für Analysen oder für maschinelles Lernen genutzt werden können. Bei einer einzelnen Rohaufnahme entsteht dabei zunächst eine ",{"type":1080,"tag":1094,"props":1523,"children":1524},{},[1525],{"type":1086,"value":1526},"Zeile mit Merkmalswerten",{"type":1086,"value":1528},". Viele solcher Zeilen zusammen bilden dann den Datensatz.",{"type":1080,"tag":1191,"props":1530,"children":1532},{"id":1531},"signal-spektrogramm-und-analyse",[1533],{"type":1086,"value":1534},"Signal, Spektrogramm und Analyse",{"type":1080,"tag":1088,"props":1536,"children":1537},{},[1538],{"type":1086,"value":1539},"Um zu verstehen, wie man aus einer Tonaufnahme Merkmale gewinnt, ist es hilfreich, das Audiosignal in verschiedenen Darstellungen zu betrachten.",{"type":1080,"tag":1088,"props":1541,"children":1542},{},[1543],{"type":1086,"value":1544},"Die folgende interaktive Darstellung zeigt ein echtes Audiosegment aus dem Datensatz. Sichtbar sind die Wellenform im Zeitbereich und das Spektrogramm als Frequenz-Zeit-Darstellung.",{"type":1080,"tag":1546,"props":1547,"children":1563},"audio-feature-extraction-visualizer",{"audio-file":1548,"title":1549,"intro":1550,"source-title":1551,"source-name":1552,"source-url":1553,"source-description":1554,"processing-note":1555,"license-name":1556,"license-url":1557,"license-note":1558,"copyright-note":1559,"window-size":1560,"hop-size":1561,"max-freq-hz":1562},"@shared/08_KuenstlicheIntelligenz/01_was_ist_ki/anuran_features_full_export/audio_segments/00000__Boana_bischoffi__INCT20955_20190911_233000.wav__0000__00.wav","Ein Froschruf als Signal und Spektrogramm","Hier siehst du ein echtes Audiosegment aus dem Datensatz. Links ist die Wellenform im Zeitbereich zu sehen, rechts das Spektrogramm auf Basis einer STFT.","Datenherkunft","AnuraSet auf Zenodo","https://doi.org/10.5281/zenodo.8342596","Die Rohaufnahmen und Annotationen stammen aus dem AnuraSet-Datensatz.","Das hier gezeigte Audiosegment wurde im Buchprojekt aus den Rohdaten abgeleitet; die betrachteten Feature-Extraktionsdaten sind selbst erzeugte, didaktisch aufbereitete Daten.","Creative Commons Attribution 1.0 Generic laut Zenodo-Eintrag","https://creativecommons.org/licenses/by/1.0/","Die AnuraSet-Projektseite und der begleitende Artikel nennen für die Daten CC0; deshalb wird hier die Quelle transparent angegeben.","Die abgeleiteten Feature-Werte und die ausgewählten Audioausschnitte wurden im Buchprojekt erzeugt. Eine eigene Weiterlizenzierung wird hier nicht festgelegt.","1024","256","5000",[],{"type":1080,"tag":1191,"props":1565,"children":1567},{"id":1566},"was-zeigt-die-darstellung",[1568],{"type":1086,"value":1569},"Was zeigt die Darstellung?",{"type":1080,"tag":1088,"props":1571,"children":1572},{},[1573],{"type":1086,"value":1574},"Die Visualisierung zeigt dasselbe Audiosignal aus verschiedenen Blickwinkeln.",{"type":1080,"tag":1576,"props":1577,"children":1579},"h3",{"id":1578},"wellenform-im-zeitbereich",[1580],{"type":1086,"value":1581},"Wellenform im Zeitbereich",{"type":1080,"tag":1088,"props":1583,"children":1584},{},[1585],{"type":1086,"value":1586},"Die Wellenform zeigt, wie sich das Signal über die Zeit verändert. Dort kann man zum Beispiel erkennen,",{"type":1080,"tag":1233,"props":1588,"children":1589},{},[1590,1595,1600],{"type":1080,"tag":1237,"props":1591,"children":1592},{},[1593],{"type":1086,"value":1594},"wann der Ruf beginnt und endet,",{"type":1080,"tag":1237,"props":1596,"children":1597},{},[1598],{"type":1086,"value":1599},"wie lange er dauert,",{"type":1080,"tag":1237,"props":1601,"children":1602},{},[1603],{"type":1086,"value":1604},"und in welchen Abschnitten das Signal besonders stark ist.",{"type":1080,"tag":1088,"props":1606,"children":1607},{},[1608],{"type":1086,"value":1609},"Die Darstellung im Zeitbereich ist besonders nützlich, wenn man die zeitliche Struktur eines Signals betrachtet.",{"type":1080,"tag":1576,"props":1611,"children":1613},{"id":1612},"spektrogramm-als-frequenz-zeit-darstellung",[1614],{"type":1086,"value":1615},"Spektrogramm als Frequenz-Zeit-Darstellung",{"type":1080,"tag":1088,"props":1617,"children":1618},{},[1619,1621,1626,1628,1633],{"type":1086,"value":1620},"Das Spektrogramm zeigt nicht nur, ",{"type":1080,"tag":1094,"props":1622,"children":1623},{},[1624],{"type":1086,"value":1625},"wann",{"type":1086,"value":1627}," etwas passiert, sondern auch, ",{"type":1080,"tag":1094,"props":1629,"children":1630},{},[1631],{"type":1086,"value":1632},"welche Frequenzen",{"type":1086,"value":1634}," in einem bestimmten Zeitabschnitt vorkommen.",{"type":1080,"tag":1088,"props":1636,"children":1637},{},[1638],{"type":1086,"value":1639},"Dazu wird das Signal in viele kurze Zeitfenster zerlegt. Für jedes dieser Fenster wird untersucht, welche Frequenzanteile darin enthalten sind. So entsteht eine Darstellung, in der Zeit und Frequenz gemeinsam sichtbar werden.",{"type":1080,"tag":1088,"props":1641,"children":1642},{},[1643],{"type":1086,"value":1644},"Helle Bereiche stehen dabei für Bereiche mit stärkerer Signalenergie.",{"type":1080,"tag":1646,"props":1647,"children":1649},"info-box",{"title":1648,"label":1001},"STFT als Analyseverfahren",[1650,1661,1666,1671],{"type":1080,"tag":1088,"props":1651,"children":1652},{},[1653,1655,1660],{"type":1086,"value":1654},"Ein wichtiges Verfahren zur Erzeugung eines Spektrogramms ist die ",{"type":1080,"tag":1094,"props":1656,"children":1657},{},[1658],{"type":1086,"value":1659},"Short-Time Fourier Transform (STFT)",{"type":1086,"value":1125},{"type":1080,"tag":1088,"props":1662,"children":1663},{},[1664],{"type":1086,"value":1665},"Dabei wird das Signal nicht als Ganzes untersucht, sondern in vielen kurzen Fenstern. Für jedes dieser Fenster wird berechnet, welche Frequenzen darin enthalten sind.",{"type":1080,"tag":1088,"props":1667,"children":1668},{},[1669],{"type":1086,"value":1670},"So kann man sehen, wie sich die Frequenzzusammensetzung des Signals über die Zeit verändert. Gerade bei Lauten, Sprache oder Tierstimmen ist das wichtig, weil sich die interessanten Strukturen im Verlauf des Signals ändern können.",{"type":1080,"tag":1088,"props":1672,"children":1673},{},[1674,1679,1681,1685],{"type":1080,"tag":1094,"props":1675,"children":1676},{},[1677],{"type":1086,"value":1678},"Definition:",{"type":1086,"value":1680}," Die ",{"type":1080,"tag":1094,"props":1682,"children":1683},{},[1684],{"type":1086,"value":1659},{"type":1086,"value":1686}," ist ein Verfahren, bei dem ein Signal in kurze Zeitfenster zerlegt wird, um für jedes Fenster die enthaltenen Frequenzen zu analysieren.",{"type":1080,"tag":1191,"props":1688,"children":1690},{"id":1689},"warum-lassen-sich-parameter-verändern",[1691],{"type":1086,"value":1692},"Warum lassen sich Parameter verändern?",{"type":1080,"tag":1088,"props":1694,"children":1695},{},[1696],{"type":1086,"value":1697},"In der Visualisierung können Parameter wie Fensterfunktion, Fensterlänge, Overlap oder DFT-Länge verändert werden.",{"type":1080,"tag":1088,"props":1699,"children":1700},{},[1701],{"type":1086,"value":1702},"Das zeigt: Eine Frequenz-Zeit-Darstellung entsteht nicht einfach automatisch, sondern hängt von Analyseentscheidungen ab.",{"type":1080,"tag":1088,"props":1704,"children":1705},{},[1706],{"type":1086,"value":1707},"Zum Beispiel gilt:",{"type":1080,"tag":1233,"props":1709,"children":1710},{},[1711,1721,1731,1741],{"type":1080,"tag":1237,"props":1712,"children":1713},{},[1714,1719],{"type":1080,"tag":1094,"props":1715,"children":1716},{},[1717],{"type":1086,"value":1718},"kleinere Fenster",{"type":1086,"value":1720}," liefern meist eine feinere Auflösung in der Zeit,",{"type":1080,"tag":1237,"props":1722,"children":1723},{},[1724,1729],{"type":1080,"tag":1094,"props":1725,"children":1726},{},[1727],{"type":1086,"value":1728},"größere Fenster",{"type":1086,"value":1730}," liefern meist eine genauere Auflösung in der Frequenz,",{"type":1080,"tag":1237,"props":1732,"children":1733},{},[1734,1739],{"type":1080,"tag":1094,"props":1735,"children":1736},{},[1737],{"type":1086,"value":1738},"Overlap",{"type":1086,"value":1740}," bestimmt, wie stark benachbarte Fenster einander überlappen,",{"type":1080,"tag":1237,"props":1742,"children":1743},{},[1744,1749],{"type":1080,"tag":1094,"props":1745,"children":1746},{},[1747],{"type":1086,"value":1748},"Zero-Padding",{"type":1086,"value":1750}," kann die Darstellung auf der Frequenzachse feiner abstufen.",{"type":1080,"tag":1088,"props":1752,"children":1753},{},[1754],{"type":1086,"value":1755},"Diese Einstellungen ändern nicht den Froschruf selbst, aber sie beeinflussen, wie das Signal analysiert und sichtbar gemacht wird.",{"type":1080,"tag":1191,"props":1757,"children":1759},{"id":1758},"wie-entstehen-daraus-daten-für-eine-tabelle",[1760],{"type":1086,"value":1761},"Wie entstehen daraus Daten für eine Tabelle?",{"type":1080,"tag":1088,"props":1763,"children":1764},{},[1765],{"type":1086,"value":1766},"Aus einer Audioaufnahme wird nicht direkt eine Tabellenzeile. Dazwischen liegen mehrere Verarbeitungsschritte.",{"type":1080,"tag":1576,"props":1768,"children":1770},{"id":1769},"typischer-ablauf",[1771],{"type":1086,"value":1772},"Typischer Ablauf",{"type":1080,"tag":1774,"props":1775,"children":1776},"ol",{},[1777,1782,1787,1792,1797,1802],{"type":1080,"tag":1237,"props":1778,"children":1779},{},[1780],{"type":1086,"value":1781},"Ein Mikrofon nimmt einen Froschruf auf und speichert ihn als Audiodatei.",{"type":1080,"tag":1237,"props":1783,"children":1784},{},[1785],{"type":1086,"value":1786},"Das Audiosignal wird digital als Folge vieler Abtastwerte dargestellt.",{"type":1080,"tag":1237,"props":1788,"children":1789},{},[1790],{"type":1086,"value":1791},"Das Signal wird analysiert, zum Beispiel mit einer STFT.",{"type":1080,"tag":1237,"props":1793,"children":1794},{},[1795],{"type":1086,"value":1796},"Aus dem Signal oder aus daraus gewonnenen Darstellungen werden beschreibende Merkmale berechnet.",{"type":1080,"tag":1237,"props":1798,"children":1799},{},[1800],{"type":1086,"value":1801},"Diese Merkmale werden als Werte in einer Tabellenzeile gespeichert.",{"type":1080,"tag":1237,"props":1803,"children":1804},{},[1805],{"type":1086,"value":1806},"Viele solcher Zeilen zusammen bilden einen Datensatz.",{"type":1080,"tag":1088,"props":1808,"children":1809},{},[1810],{"type":1086,"value":1811},"Man kann den Ablauf auch so zusammenfassen:",{"type":1080,"tag":1813,"props":1814,"children":1815},"table",{},[1816,1835],{"type":1080,"tag":1817,"props":1818,"children":1819},"thead",{},[1820],{"type":1080,"tag":1821,"props":1822,"children":1823},"tr",{},[1824,1830],{"type":1080,"tag":1825,"props":1826,"children":1827},"th",{},[1828],{"type":1086,"value":1829},"Stufe",{"type":1080,"tag":1825,"props":1831,"children":1832},{},[1833],{"type":1086,"value":1834},"Beispiel",{"type":1080,"tag":1836,"props":1837,"children":1838},"tbody",{},[1839,1858,1871,1884,1924],{"type":1080,"tag":1821,"props":1840,"children":1841},{},[1842,1847],{"type":1080,"tag":1843,"props":1844,"children":1845},"td",{},[1846],{"type":1086,"value":1145},{"type":1080,"tag":1843,"props":1848,"children":1849},{},[1850,1856],{"type":1080,"tag":1102,"props":1851,"children":1853},{"className":1852},[],[1854],{"type":1086,"value":1855},".wav",{"type":1086,"value":1857},"-Datei mit einem Froschruf",{"type":1080,"tag":1821,"props":1859,"children":1860},{},[1861,1866],{"type":1080,"tag":1843,"props":1862,"children":1863},{},[1864],{"type":1086,"value":1865},"Signal im Zeitbereich",{"type":1080,"tag":1843,"props":1867,"children":1868},{},[1869],{"type":1086,"value":1870},"Folge von Amplitudenwerten",{"type":1080,"tag":1821,"props":1872,"children":1873},{},[1874,1879],{"type":1080,"tag":1843,"props":1875,"children":1876},{},[1877],{"type":1086,"value":1878},"Frequenz-Zeit-Darstellung",{"type":1080,"tag":1843,"props":1880,"children":1881},{},[1882],{"type":1086,"value":1883},"Spektrogramm aus einer STFT",{"type":1080,"tag":1821,"props":1885,"children":1886},{},[1887,1892],{"type":1080,"tag":1843,"props":1888,"children":1889},{},[1890],{"type":1086,"value":1891},"Extrahierte Merkmale",{"type":1080,"tag":1843,"props":1893,"children":1894},{},[1895,1900,1901,1906,1907,1912,1913,1918,1919],{"type":1080,"tag":1102,"props":1896,"children":1898},{"className":1897},[],[1899],{"type":1086,"value":1107},{"type":1086,"value":1109},{"type":1080,"tag":1102,"props":1902,"children":1904},{"className":1903},[],[1905],{"type":1086,"value":1461},{"type":1086,"value":1109},{"type":1080,"tag":1102,"props":1908,"children":1910},{"className":1909},[],[1911],{"type":1086,"value":1123},{"type":1086,"value":1109},{"type":1080,"tag":1102,"props":1914,"children":1916},{"className":1915},[],[1917],{"type":1086,"value":1115},{"type":1086,"value":1109},{"type":1080,"tag":1102,"props":1920,"children":1922},{"className":1921},[],[1923],{"type":1086,"value":1501},{"type":1080,"tag":1821,"props":1925,"children":1926},{},[1927,1932],{"type":1080,"tag":1843,"props":1928,"children":1929},{},[1930],{"type":1086,"value":1931},"Datensatz",{"type":1080,"tag":1843,"props":1933,"children":1934},{},[1935],{"type":1086,"value":1936},"eine Tabellenzeile mit diesen Merkmalswerten",{"type":1080,"tag":1576,"props":1938,"children":1940},{"id":1939},"merksatz",[1941],{"type":1086,"value":1942},"Merksatz",{"type":1080,"tag":1165,"props":1944,"children":1945},{},[1946],{"type":1080,"tag":1088,"props":1947,"children":1948},{},[1949],{"type":1086,"value":1950},"Feature Extraction verbindet die Rohdaten mit dem späteren Datensatz.\nAus einer Aufnahme werden Merkmalswerte, aus vielen Merkmalszeilen wird eine Lerntabelle.",{"type":1080,"tag":1191,"props":1952,"children":1954},{"id":1953},"warum-ist-feature-extraction-wichtig",[1955],{"type":1086,"value":1956},"Warum ist Feature Extraction wichtig?",{"type":1080,"tag":1088,"props":1958,"children":1959},{},[1960],{"type":1086,"value":1961},"Viele Lernverfahren arbeiten besser, wenn die wichtigsten Eigenschaften eines Signals bereits in kompakter Form beschrieben sind.",{"type":1080,"tag":1088,"props":1963,"children":1964},{},[1965],{"type":1086,"value":1966},"Das hat mehrere Vorteile:",{"type":1080,"tag":1233,"props":1968,"children":1969},{},[1970,1981,1992,2003,2014],{"type":1080,"tag":1237,"props":1971,"children":1972},{},[1973,1975,1980],{"type":1086,"value":1974},"Die Daten werden ",{"type":1080,"tag":1094,"props":1976,"children":1977},{},[1978],{"type":1086,"value":1979},"kompakter",{"type":1086,"value":1125},{"type":1080,"tag":1237,"props":1982,"children":1983},{},[1984,1986,1991],{"type":1086,"value":1985},"Verschiedene Aufnahmen werden ",{"type":1080,"tag":1094,"props":1987,"children":1988},{},[1989],{"type":1086,"value":1990},"leichter vergleichbar",{"type":1086,"value":1125},{"type":1080,"tag":1237,"props":1993,"children":1994},{},[1995,1997,2002],{"type":1086,"value":1996},"Wichtige Eigenschaften werden ",{"type":1080,"tag":1094,"props":1998,"children":1999},{},[2000],{"type":1086,"value":2001},"gezielt beschrieben",{"type":1086,"value":1125},{"type":1080,"tag":1237,"props":2004,"children":2005},{},[2006,2008,2013],{"type":1086,"value":2007},"Aus vielen einzelnen Aufnahmen entsteht ein ",{"type":1080,"tag":1094,"props":2009,"children":2010},{},[2011],{"type":1086,"value":2012},"einheitlicher Datensatz",{"type":1086,"value":1125},{"type":1080,"tag":1237,"props":2015,"children":2016},{},[2017],{"type":1086,"value":2018},"Auf diesem Datensatz kann man später Verfahren des maschinellen Lernens anwenden.",{"type":1080,"tag":1088,"props":2020,"children":2021},{},[2022],{"type":1086,"value":2023},"Im Beispiel der Froschrufe kann ein Modell später lernen, aus den extrahierten Merkmalen vorherzusagen, zu welcher Froschart ein neues Audiosegment gehört.",{"type":1080,"tag":1191,"props":2025,"children":2027},{"id":2026},"einordnung",[2028],{"type":1086,"value":2029},"Einordnung",{"type":1080,"tag":1088,"props":2031,"children":2032},{},[2033],{"type":1086,"value":2034},"Die Grundidee der Feature Extraction gilt nicht nur für Audiodaten.",{"type":1080,"tag":1088,"props":2036,"children":2037},{},[2038],{"type":1086,"value":2039},"Auch bei anderen Datenarten werden oft zuerst Merkmale gewonnen:",{"type":1080,"tag":1233,"props":2041,"children":2042},{},[2043,2053,2063],{"type":1080,"tag":1237,"props":2044,"children":2045},{},[2046,2047,2051],{"type":1086,"value":1325},{"type":1080,"tag":1094,"props":2048,"children":2049},{},[2050],{"type":1086,"value":1330},{"type":1086,"value":2052}," zum Beispiel Wortanzahlen oder Häufigkeiten",{"type":1080,"tag":1237,"props":2054,"children":2055},{},[2056,2057,2061],{"type":1086,"value":1325},{"type":1080,"tag":1094,"props":2058,"children":2059},{},[2060],{"type":1086,"value":1341},{"type":1086,"value":2062}," zum Beispiel Farben, Kanten oder Formen",{"type":1080,"tag":1237,"props":2064,"children":2065},{},[2066,2067,2071],{"type":1086,"value":1325},{"type":1080,"tag":1094,"props":2068,"children":2069},{},[2070],{"type":1086,"value":1352},{"type":1086,"value":2072}," zum Beispiel Mittelwerte, Trends oder charakteristische Schwankungen",{"type":1080,"tag":1088,"props":2074,"children":2075},{},[2076],{"type":1086,"value":2077},"Das Audio-Beispiel zeigt also einen speziellen Fall eines allgemeinen Prinzips:",{"type":1080,"tag":1088,"props":2079,"children":2080},{},[2081],{"type":1086,"value":2082},"Rohdaten werden verarbeitet, daraus entstehen Merkmale, und diese Merkmale bilden später die Grundlage für einen Datensatz und für maschinelles Lernen.",{"type":1080,"tag":1191,"props":2084,"children":2086},{"id":2085},"zusammenfassung",[2087],{"type":1086,"value":2088},"Zusammenfassung",{"type":1080,"tag":1576,"props":2090,"children":2092},{"id":2091},"merksatz-1",[2093],{"type":1086,"value":1942},{"type":1080,"tag":1165,"props":2095,"children":2096},{},[2097],{"type":1080,"tag":1088,"props":2098,"children":2099},{},[2100,2104,2105,2109,2110,2114],{"type":1080,"tag":1094,"props":2101,"children":2102},{},[2103],{"type":1086,"value":326},{"type":1086,"value":1176},{"type":1080,"tag":1094,"props":2106,"children":2107},{},[2108],{"type":1086,"value":1181},{"type":1086,"value":1183},{"type":1080,"tag":1094,"props":2111,"children":2112},{},[2113],{"type":1086,"value":1188},{"type":1086,"value":1125},{"type":1080,"tag":1088,"props":2116,"children":2117},{},[2118],{"type":1086,"value":2119},"Am Beispiel der Froschrufe sieht man diesen Übergang besonders gut:",{"type":1080,"tag":1233,"props":2121,"children":2122},{},[2123,2128,2133,2138,2143],{"type":1080,"tag":1237,"props":2124,"children":2125},{},[2126],{"type":1086,"value":2127},"Die Rohdaten sind Audioaufnahmen.",{"type":1080,"tag":1237,"props":2129,"children":2130},{},[2131],{"type":1086,"value":2132},"Daraus wird ein digitales Signal.",{"type":1080,"tag":1237,"props":2134,"children":2135},{},[2136],{"type":1086,"value":2137},"Mit Verfahren wie der STFT lassen sich wichtige Strukturen des Signals untersuchen.",{"type":1080,"tag":1237,"props":2139,"children":2140},{},[2141],{"type":1086,"value":2142},"Anschließend werden Merkmale wie Dauer, Energie oder dominante Frequenz berechnet.",{"type":1080,"tag":1237,"props":2144,"children":2145},{},[2146],{"type":1086,"value":2147},"Diese Merkmale werden als Spalten eines Datensatzes gespeichert.",{"type":1080,"tag":1088,"props":2149,"children":2150},{},[2151],{"type":1086,"value":2152},"So wird aus einer Rohaufnahme eine Form von Daten, mit der man statistisch arbeiten und später ein Modell trainieren kann.",{"title":62,"searchDepth":33,"depth":33,"links":2154},[2155,2156,2157,2158,2159,2160,2161,2165,2166,2170,2171,2172],{"id":1193,"depth":33,"text":1196},{"id":1223,"depth":33,"text":1226},{"id":1309,"depth":33,"text":1312},{"id":1379,"depth":33,"text":1382},{"id":1427,"depth":33,"text":1430},{"id":1531,"depth":33,"text":1534},{"id":1566,"depth":33,"text":1569,"children":2162},[2163,2164],{"id":1578,"depth":59,"text":1581},{"id":1612,"depth":59,"text":1615},{"id":1689,"depth":33,"text":1692},{"id":1758,"depth":33,"text":1761,"children":2167},[2168,2169],{"id":1769,"depth":59,"text":1772},{"id":1939,"depth":59,"text":1942},{"id":1953,"depth":33,"text":1956},{"id":2026,"depth":33,"text":2029},{"id":2085,"depth":33,"text":2088,"children":2173},[2174],{"id":2091,"depth":59,"text":1942},"markdown","content:08_KuenstlicheIntelligenz:01_was_ist_ki:04_featureextraction:de.md","08_KuenstlicheIntelligenz/01_was_ist_ki/04_featureextraction/de.md","08_KuenstlicheIntelligenz/01_was_ist_ki/04_featureextraction/de","md",1787079994799]