New Machine Learning Model to Movie Recommender and Sentiment Analysis

Recently, technology has put the limelight on every industry, which ends in an overload of massive data that might be resolved using efficient recommender systems. The recommendation system is handy to accommodate large amounts of knowledge and separate helpful information. One in each of its applications falls in movie recommendations. Currently, numerous movies are developing, which makes it difficult for movie addicts to settle on one in all of them. The movie recommendation system intends to assist movie enthusiasts in cutting back on human effort by suggesting movies efficiently and effectively without wasting much time in useless browsing. Recommending products that support user interest is one of the principal business tasks achieved using machine learning (ML). This research develops a new supervised ML model to recommend movies with the naïve Bayes’ (NB) and support vector machine (SVM) classifier that combines cosine similarity (CS) and sentiment analysis (SA). The CS may be a standard accustomed measure of finding similarity between two items, and the SA of a movie review can classify the type of a review. The proposed model applies the training strategy and the reviews classifications are done accordingly and the predictions are made. This research model performance is tested with benchmark instances, 97% accuracy is obtained by applying SA with the NB approach, and 98% accuracy is obtained using an SVM. The simulation results conclude that the proposed model produced more significant outcomes than the existing approaches in deploying the proposed recommender engine as a website that takes input from users and recommends similar movies with related images in minimal computing time.