Proposing a Meta-Heuristic Approach for the Long Tail Problem of Recommender Systems

Hosseini, Sayedeh Zahra; Mohammadi, Azadeh

Abstract

Recommendation systems are a solution for providing appropriate suggestions to users and helping them in the decision-making process. In the most recommendation systems, the purpose is to offer items tailored to the user's interests based on the past scores. Focusing on the previous scores causes various problems including ignoring long tail items. Long tail items are those that are rated by a small number of users, therefor they are often not recommended by recommendation systems. This leads to the recommendation system bias towards offering previously popular items and ignoring the diversity and novelty of suggestions. To solve this problem, in this paper, a multi-objective optimization approach is used to increase the accuracy and chance of recommending long tail items. In the proposed method which is applied to the Movielens dataset, first, various groups of items are created by categorizing movies based on freshness (year of production). Then by applying the NSGAII algorithm, accuracy and diversity of suggested items are optimized. The results of applying the proposed method to the mentioned dataset indicate that the suggested method, while maintaining the accuracy of recommendation, has been able to increase the number of long tail items in the recommendation lists.

Citation

Hosseini, S. Z., & Mohammadi, A. (2021). Proposing a Meta-Heuristic Approach for the Long Tail Problem of Recommender Systems. . https://doi.org/10.1109/ICWR51868.2021

• Conference Name: 7th International Conference on Web Research (ICWR)
• Conference Location: Tehran - Iran
• Start Date: May 19, 2021
• End Date: May 20, 2021
• Acceptance Date: Feb 19, 2021
• Online Publication Date: May 20, 2021
• Publication Date: May 20, 2021
• Deposit Date: Sep 25, 2023
• Publisher: Institute of Electrical and Electronics Engineers
• DOI: https://doi.org/10.1109/ICWR51868.2021
• Keywords: Recommender System; Long Tail Items; Freshness; Multi Objective Optimization; Collaborative Filtering; NSGAII