A Recommendation Algorithm Incorporating Moth-Flame Optimization Algorithm and Fuzzy Clustering
Xueli Shen, Yihui Sun · 2023
Aiming at the issues of data sparsity, restrictions on discovering nearest neighbors, and accuracy of similarity calculation of conventional collaborative filtering techniques, a hybrid recommendation algorithm based on MFO-optimized fuzzy clustering and combined similarity is developed. The computation of similarity for the differences in feature attributes of various users is done using a multi-attribute similarity calculation model that incorporates user feature attributes. Based on the different fuzzy clustering affiliation levels, the nearby neighbors are filtered, the mothballing method (MFO) is included, the initialized clustering center is identified, and rating prediction of the clustered results is carried out. The experimental results on the MovieLens dataset demonstrate that the algorithm successfully addresses the issue of data sparsity, dramatically lowers the recommendation error and provides a more thorough examination of the recommendation outcomes than the conventional approaches.