Amazon Review Helpfulness Prediction
In this project, exploratory analysis of relationship between helpfulness votes received by a review and various features was performed. Based on analysis, I designed an Elastic Net Regressor to predict helpfulness votes received by a review and acheived a Mean Absolute Error of 0.16705956391.
This project was developed as a part of UCSD Kaggle contest and finished among top 5% in leaderboard.