A Comprehensive Trust-Based Item Evaluation Model for Recommendation in Social Network

作者：Cui Laizhong^*; Ou Peng; Lu Nan; Zhang Guanjing

来源：IEEE Symposium on Computers and Communication (ISCC), Messina, Italy, 2016-06-27 To 2016-07-01.

DOI：10.1109/ISCC.2016.7543881

• Collaborative filtering algorithms
• Evaluation modeling
• Evaluation results
• On-line social networks
• Trust computations

With the rapid development of online social network, people tend to express opinions and obtain information in social network. Due to the overwhelming amount of data in social network, users resort to recommendation system to find appropriate services or items. However, most current recommendation systems rely on the user to calculate the item evaluation results, which has many limitations in the item evaluation. To solve these limitations on item evaluation in social network, we put forward a new trust-based item evaluation model (called CTDR), which integrates trust, domain inclination and item reputation in this paper. First, we use the similarity between users, interaction, as well as the reputation of the target user to introduce a new trust computation method. Second, we introduce the concept of domain inclination and combine the domain activity of user and domain popularity to calculate the domain inclination. Third, we make use of the favorable rate of an item to compute the item reputation. Through extensive experiments, we evaluate CTDR with the traditional collaborative filtering algorithm. The experimental results validate the effectiveness and performance of our trust-based item evaluation model.

• 出版日期2016
• 单位深圳大学