Application of Enhanced Memory-Based Collaborative Filtering Algorithm to a Tourist Spot Recommender System

Abstract

Recommender systems use collaborative filtering, where information is filtered by using the recommendations from different people. A tourist spot recommender system is developed in this paper using the enhanced memory-based collaborative filtering algorithm based on user similarity. The enhancement utilizes a new similarity measure that was devised to identify co-rated items and computes the user similarity. The performance of the enhanced algorithm was evaluated using the standard evaluation metrics, and the accuracy was compared with the traditional Cosine, Euclidean Distance, and Pearson Correlation similarity metrics. The application of the enhanced algorithm in a tourist spot recommender system validated the model in providing accurate recommendations to similar users who previously rated the tourist spots. The superior performance and accuracy exhibited by the recommender system that uses the new similarity measure formulated in this study showed that it is an effective solution to improve the identification of co-rated items and user similarity.