search

nkartik94 / Multi-Label-Text-Classification

Kaggle Toxic Comments Challenge
106 stars 54 forks source link

why do OneVsRestClassifier for each topic, and not all at once? #4

Open yishairasowsky opened 4 years ago

yishairasowsky commented 4 years ago

hi @nkartik94 ! great presenetation! i love your project! so very informative to beginners like me! is there a reason you can exaplain why in OneVsRestClassifier you trained the model for each category instead of just training the model once and for all? thanks!!

SwaggerlisiousSteven commented 3 years ago

hi @nkartik94 ! great presenetation! i love your project! so very informative to beginners like me! is there a reason you can exaplain why in OneVsRestClassifier you trained the model for each category instead of just training the model once and for all? thanks!!

The reason why he used OneVsRest is because logistical regression cannot identify multilabels, algorithms such as RandomForest or Neural Networks can.