Closed yujinjcho closed 7 years ago
So I used another group of the part_files and looks like lambda 10 is working the best
This is pretty kewl. susan-granger gives fresh rating ~57% of time. Model ~80%+ correct. Also, some pretty high lambdas
Interesting.. so does this mean it's just severely overfitting the training set? Because of how many features we have?
I think so
lol check this out
Section I had for cv set
Section I had for test set