Movie Recommender System Based on K-Means Dynamic Collaborative Filtering

Abstract

In today's world, the internet has become an indispensable tool for everyday life. Machine learning models that learn from themselves without being explicitly programmed are becoming increasingly important as the internet grows. Recommendation plays a crucial role in thisworld of computers. One can get lost in a sea of information without guidance. Movies are a popular medium of entertainment as well. Most people watch movies that are suggested by others. Every individual likes different type of movies. So a movie recommender system can apparantly increase viewers count for the movies and also reduces time for searching the movies. Traditional movie recommendation system are based on either Content Based (CB) Filtering technique or Collaborative Filtering (CF) technique that produces result in a minimum number. A combination of the Content Based (CB) and Collaborative Filtering (CF) approach is used here to recommend the most similar movies based on the user’s choice. A K-means Dynamic Collaborative Filtering (DCF) approach is used to predict the most relevant recommendation for the user. Many websites provide dataset for the research purpose on machine learning models. Here, MovieLens dataset is used which contains millions of data of movie along with user ratings.