Certain Investigations on Cognitive based Movie Recommendation system using Pairwise Cosine Similarity

Abstract

The Internet serves as a major source in various sectors to communicate and share data. It is being used by every individual across the world for several purposes. The main objective of the recommender system is to provide users with suitable products and information. The movie recommendation system also helps in providing customized decisions to the user, thus improving the experience of customer or user. Content-based filtering manipulates item attributes to suggest other items that are relevant to user likes, this suggestion is made using past actions of customers or users. Cosine Similarity is an evaluation metric that helps in finding how similar the items are regardless of their dimensions. The proposed study, furnish users with suggestions that rest on the prevalence and category of a movie. Content based recommendation system is built using cosine similarity. The movie recommendation system has its separate algorithms for making suggestions by examining users’ past data. Many attributes are considered like rating of each movie, name of the director and genre of the film. The system make recommendation by considering one attribute or mixture of attributes. For the proposed system, two different datasets used namely, the credits dataset and the movie dataset. The analysis of data is done with help of Python.