Outlier Analysis a pre processing technique

Do we need to find the outlier for the dependent variable or its only meant to be found for independent variable or for both.

As per my understanding , outliers should be removed from independent variables as our model needs to be generalized basically and any extreme value must not affect it as such .Also you may refer to this blog for a deeper insight : https://heartbeat.fritz.ai/how-to-make-your-machine-learning-models-robust-to-outliers-44d404067d07

1 Like

Thanks for the reply can you tell me about the scaling of the dependent variables is valid or not?

Hi swarup , the main aim behind scaling any variable is increase in the computational speed .We may scale dependent variables but while predicting make sure to convert back in the same manner .

© Copyright 2013-2019 Analytics Vidhya