Effective user preference mining-based personalised movie recommendation system

One of the primary issues of many websites is the suggestion of multiple choices to the users at the same time, which makes the task more complex and time consuming to find the right product. Web mining and recommendation system based on user behaviour helps users by providing essential information without asking explicitly. Several movie recommendation systems are available to suggest movies, but often they do not do that effectively. To achieve enhanced effectiveness and efficiency, user's movie ratings were retrieved, cleaned, formatted and grouped into proper, meaningful session and data profile was developed. In this paper, we have developed a new ontology for clear and better understanding of the movie domain. The user data consisting of movie ratings is used to recommend movies for the users. For the classification of users, we use adaptive K-nearest neighbour (AKNN) approach and post classification process; movies are recommended to the active target user. The obtained results of the proposed recommendation approach are compared with existing baseline methods, and the results prove that the presented approach to be proficient.