The algorithms described in the report are implemented in different file.
For example, "train_svm" implements kernerlized SVM.

Training data and testing data is in the data/ folder. The filename is formatted as 
"feature - tweets num - vector size"
For example, "W2Vtrain2000-300" means the feature vector is generated by Word2Vec, using 2000 tweets per movie and vector size of 300.