Investigate manually selected descriptors of ligands

[abduskhazi]

Investe manually selected 121 of the ligand descriptors. Continue to use all of the protein descriptors.

ligand_descriptors.ods

[abduskhazi]

priority 1

[abduskhazi]

Hi All, I completed the implementation of manual feature selection. With this I am getting the following outputs -

• Linear Regression : Training R2 score = 0.365 & Validation R2 score = 0.342 (Execution ID = 4286703491)
• Random Forest Regressor : oob score = 0.744 & Validation R2 score = 0.747 (Execution ID = 2911216629)

These results are less accurate as compared to the models using all the features.

For running this you need to call the following function to get the data

import data_bakery as bakery
X, y, features = bakery.bake_train_Xy_manual_feature_selection()

Regards, Abdus Salam Khazi