Machine learning based recommendation system on movie reviews using KNN classifiers

Abstract

Recommender systems are the systems that are designed to recommend items to the consumer depending on several different criteria. These systems estimate the most possible product that the consumers are most likely to buy and are of interest to. Companies like Netflix, Amazon, etc. use recommender services to allow their customers to find the right items or movies for them.In the current system recommendations, the content of ltering and collective ltering typically fall into two groups. The method is formerly Periment in our paper in all methods. We take film features such as stars, directors, for content-based ltering. Movie definition and keywords as inputs use TF-IDF and doc2vec for measuring the film resemblance. For the first time, Input to our algorithm is the film ranking encountered by users, and we use neighbours nearest K, as Factorization of matrix to estimate film scores for consumers. We find that teamwork functions better than content. Predictive error and estimation time ltering.