search

amritbhanu / EDM591_Hyperparameter

Hyper Parameter Optimization for Education Mining
0 stars 1 forks source link

Feature selection #12

Open amritbhanu opened 6 years ago

amritbhanu commented 6 years ago

Use baseline to find the best features, which is the most important thing affect students. Is the result reliable? Apply haperparameter find the best features, compared with baseline. (find the best for each cross validation or the entire????)(save the temp results)(find best for each and combine with occur frequency)

Dataset 1: features for student grade Dataset 2: features for social feature influence student grade Dateset 3: features for group project grade

amritbhanu commented 6 years ago

Problem, with SVM we can not get feature importance for other kernels except Linear kernel