A Novel Approach for Product Recommendation using XGBOOST

Abstract

Sentiment analysis is the most trending research area. Generally, most purchase decisions and price predictions are made based on product reviews. Sentiment analysis helps in understanding the product better. The sentiment analysis of a product summarizes whether the product has a positive, negative or neutral rating. Existing machine learning algorithms like logistic Regression, Decision Tree are used to determine sentiment for product reviews. This work includes XGBOOST and a hybrid model XGBOOST - RF used to observe sentiment on product reviews. The model that gives best performance is used to build a system that recommends products to users.