Evaluation of Group Modelling Strategy in Model-Based Collaborative Filtering Recommendation
Rosmamalmi Mat Nawi, Shahrul Azman Mohd Noah, Lailatul Qadri Zakaria · International Journal of Machine Learning and Computing · 2020
Recommender systems for groups are becoming increasingly popular since many information needs instigate from group and social activities, such as listening to music, watching movies, and traveling.One of the important aspects in group recommendation is group modelling aggregation strategy which is a process to generate the overall ratings of the group.Such ratings are considered as representations of the groups.There are few group aggregations approaches.In this paper we evaluated two group aggregation approaches which are the Most Pleasure and Average strategy group modelling.We implemented both approaches on the model-based collaborative filtering technique using the single value decomposition and average least square prediction algorithms.The experimental results show that the Average strategy outperformed the Most Pleasure strategy for both prediction algorithms in terms of MAE, RMSE, and precision and recall metrics.