A Hybrid Movie Recommendation System using Particle Swarm Optimization and K-means Clustering Algorithm

Abstract

On the Internet, where the number of choices is overwhelming, priorities need to be prioritized and relevant information provided efficiently to make selection easier and to minimize the problem of information overload, which has created a potential problem for many Internet users. Recommender systems solve this problem by searching through large amounts of dynamically generated information to provide users with personalized content and services. This essentially makes recommendation systems a central component of websites and e-commerce applications. Nowadays, cold-start and scalability problems are the researchable areas of recommendation systems. Solving these two problems will result in personalized recommendation and increased recommendation accuracy. To solve these problems, this research paper focuses on movie recommendation systems, the main goal of which is to propose a recommendation system through data clustering and computational intelligence. In this research article a novel recommendation system was discussed, which is a Hybrid Particle Swarm Optimization (PSO) and k-means algorithm (PSO-KM) applied to the MovieLens (ml-latest-small) dataset. This dataset first passed through the successive stages of Pre-processing process and produced quality and numerical data. After the quality and numerical data is produced, the Elbow Method determined the number of clusters or K as initial seeds of K-means clustering algorithm. Then, the Hybrid PSO-KM algorithm runs on the dataset and produced compact and homogenous clusters. The result of evaluation matrices show that the proposed Hybrid algorithm produced 0.93% more compact clusters and 0.97% better separation between neighboring clusters than the k-means clustering algorithm.