Fake News Detection on Liar dataset
for preprecessing the metadata and the text data.
Processing the metadata
contains utility functions for term doc matrix, tfidf vectorizer, POS tagging, glove model and confusion matrix to be used in the other files
contains classifiers like lienar svm, svm with sgd, naive-bayes, random forest and logistic regression for classifying the data
contains the code for cnn model implementation on the text data.
- process metadata
- process text data
- XGBoost on metadata
- CNN model
- LSTM model
- ensemble of best models