Creating 2 models to improve accuracy

Hi am building a classification model

I have trained a model and I have taken the misclassified output and again given it to another model which is giving me best accuracy at the end

How to pipeline these 2 models

Is my approach right

@raviteja1993 Without knowing the specifics of the models you have mentioned, you could create an ensemble of them, where the final prediction will be some weighted average/mode of their individual predictions.

@raj56 I am working on the ensemble approach of two models can you please provide some good material for a better understanding of the ensemble approach.


@raviteja1993 I do not know what library/language you are using.
Scikit learning has good support for ensembles.
Kindly refer to this -

Analytics Vidhya itself has a good resource

I believe Boosting algorithms would be more suitable to your use-case, where one model rectifies the errors made by the previous model

© Copyright 2013-2020 Analytics Vidhya