An effective recommendation method for cold start new users using trust and distrust networks

  • Authors:

  • Chien Chin Chen;Yu-Hao Wan;Meng-Chieh Chung;Yu-Chun Sun

  • Affiliations:

  • Department of Information Management, National Taiwan University, Taiwan;Department of Information Management, National Taiwan University, Taiwan;Department of Information Management, National Taiwan University, Taiwan;Department of Information Management, National Taiwan University, Taiwan

  • Venue:
  • Information Sciences: an International Journal
  • Year:
  • 2013

Quantified Score

Hi-index 0.07

Visualization

Abstract

Recommendation systems analyze the purchasing behavior (e.g., item ratings) of users to learn about their preferences and recommend products or services that may be of interest to them. However, as new users require time to become familiar with recommendation systems, the systems usually have limited information about newcomers and have difficulty providing appropriate recommendations. This so-called new user cold start phenomenon has a serious impact on the performance of recommendation systems. As a result, there has been increasing research in recent years into new user cold start recommendation methods that try to provide useful item recommendations for cold start new users. The rationale behind much of the research is that recommending items to new users generally creates a sense of belonging and loyalty, and encourages them to frequently utilize recommendation systems. In this paper, we propose a cold start recommendation method for the new user that integrates a user model with trust and distrust networks to identify trustworthy users. The suggestions of these users are then aggregated to provide useful recommendations for cold start new users. Experiments based on the well-known Epinions dataset demonstrate the efficacy of the proposed method. Moreover, the method outperforms well-known recommendation methods for cold start new users in terms of the recall rate, F1 score, coverage rate, users coverage, and execution time, without a significant reduction in the precision of the recommendations.