Prediction of tweets topics (sports or politics)
In this project, I followed two approaches to solve this classification problem.
The first one is by using the tf-idf technique following by applying Multinomial Naive Bayes and SVM modeles.
The second one is by using Word Embedding (pre trained model) and feed them to a LSTM classifier.
TF-IDF + SVM results were better than Word2Vec + LSTM (0.96 vs 0.94).