Classification of customer review using random forest classifier

Customer review has evolved into an indicator of a person’s judgment in the decision-making process for a specific entity. The growing amount of review data on the Internet provides numerous opportunities for people to find important information. Reviews are very important in business because they allow business people to assess the level of consumer interest in a product that is about to be released. Sentiment analysis can help you solve the problem of categorizing reviews as positive or negative. The purpose of the article is to combine a sentiment analysis technique with a machine learning approach. The Random Forest Classifier is used to classify sentiment groups, which improves sentiment analysis performance significantly. Implementing an imbalanced SMOTE technique improves model performance during preprocessing. The results show that Random Forest on the electronic review can achieve an accuracy of 92 % in product review classification.