Quoa Question Pair Similarity Project

Techniques Used:

1. BOW
2. TFIDF
3. Own Word Embedding
4. Pre-Trained Word2Vec
5. Pre-Trained GloVe
6. Pre-Trained BERT
7. Senetence Similarity

Extracted Features:

1. Basic Word features
2. Length Based Features
3. Token Based Features
4. Fuzzy features
5. Cosine similarity between two sentences

Machine Learning Models Used:

1. Logistic Regression
2. SVM
3. Naive Bayes
4. KNN
5. Decision Tree
6. Random Forest
7. Gradient Boost
8. ADA Boost
9. XGBoost

Best Method:

XGBoost using BERT and Advanced Extracted features Accuracy of 86%.