Restricted Boltzmann Machine and Matrix Factorization-Alternating Square Algorithm for Development Tourist Recommendation System

Abstract

Abstract The rapid development of technology affects the growth of tourist attraction information in Indonesia. Therefore, an accurate recommendation system is needed in recommending tourist attractions. In this final project, we use the Collaborative Filtering method, namely the Restricted Boltzmann Machine (RBM) algorithm and the Matrix Factorization-Alternating Least Squares (MF-ALS) algorithm in recommending tourist attractions. Attraction recommendations will be generated from the type of tourist attraction available on the website and the rating that has been given by previous users who have visited the tourist attraction. We use a root mean square error (RMSE) to find the accuracy. From the results of the research and implementation of the two algorithms, it can be concluded that the RBM algorithm is more accurate than the MF-ALS algorithm. The RBM algorithm has an RMSE value of 41%, while the MF-ALS algorithm 81%.