Best explanation I found on Lasso code! Thanks for making this! :)

Very well explained..

excellent Sir, will it be possible to generate a pipeline to feed in the selected features to the regression or classification model to perform model evaluation and analytics - Thank yiou

@yourdatateacher1776

Жыл бұрын

Yes, sure. You should use the SleectFromModel object in scikit learn, that selects the most important features selected by a previous model and allows you to feed another model with them.

@KN-tx7sd

Жыл бұрын

@@yourdatateacher1776 Sir, thank you

i think this video is complete, as we have list of imp feature then we need to drop others non imp. feature from dataframe.

@yourdatateacher1776

2 жыл бұрын

Thanks! I'm glad you liked it!

@dewman7477

2 жыл бұрын

@@yourdatateacher1776 What are the other methods of feature selection? How do I normally find optimal number of features for my models?

@yourdatateacher1776

2 жыл бұрын

@@dewman7477 other methods of feature selection are filter-based feature selection and Recursive Feature Elimination. I prefer the latter because it's related to the model logic. The former can be used to remove the less relevant features when your dataset has too many columns.

@beautyisinmind2163

2 жыл бұрын

if we have multi class classification problem then can we use lasso and ridge for feature selection? or it is only suitable for binary problem only? I mean Iris data set it is a multi class classification problem so can we apply lasso or ridge for it to select features?

@yourdatateacher1776

Жыл бұрын

@@beautyisinmind2163 with multiclass classification you need to use logistic regression with l1 or l2 penalty functions.