A Recommended System based on MovieLens data.
-cfn: add collaborative features (optional) -contn: use SVD to decompose content features (optional) N: the number of columns for latent collaborative features after SVD M: the number of columns for content features after SVD
command line: python training.py [-cfn N] [contn M]
table: rf_max_depth rf_n cf_n cf_filter accuracy 2 20 NA 0.70297 2 20 20 70 0.70280 2 20 200 70 0.70170 5 20 100 55 0.70080