Tutorial 88 - Feature selection for speeding up machine learning training​

This video goes through the process of extracting features from training images, balancing data between multiple classes, fitting a Random Forest model, then finding useful features using Boruta followed by fitting a model using only the select subset of features.

Boruta library information:

pip install Boruta

[familywu]

Thank you very much for this excellent course, Professor. I have learned a lot from it.

[PEEYUSHKP]

Sir I have tried this code in my application. But I was not able to execute Canny and Roberts feature extraction sections. These two parts alone is throwing errors. Can you help sir.

[PEEYUSHKP]

Can we use this feature selection for classification also sir. How can we do it sir.

[Atfatra]

Thank you for your work. It taught me a lot.

A question: why we do not normalize data? Like in Tutorial 66a. It's not needed in this case or it's just optional?

[mohamedtouati]

Sir, I have a shaded data problem and my treatment for diabetic retinopathy, I want to solve the detection and classification problem, but its model is too small with different sizes. whats your suggessions ?

[kabilakamal]

Very educative indeed. Thanks for your awesome job🙏🏻

Please how do I contact you?