Personalized Trusted Service Recommendation Method based on Social Work

다국어 초록 (Multilingual Abstract)

In order to improve credibility and precision of the personalized service recommendation, we propose a personalized trusted services recommendation method based on social network. Firstly, according to the similarity of user's preference, the method build new relations combined with direct trust and indirect trust of users between them. Secondly, form a set of similar services which includes services invoked by target users. Eventually, get user’s indirect trustrank to the recommended services through the weighted average of the target users’ indirect trustrank to the recommended service and thus form a recommendation list. The experiment take precision rate, recall rate and F1_Measure as evaluation indexes. Comparing with traditional method which based on no confidence and users’ trusted network, this approach demonstrates its effectiveness and feasibility through the result of the experiment.

목차 (Table of Contents)

• Abstract
• 1. Introduction
• 2. Background and Related Work
• 3. The Trust Relationship
• 3.1. Direct Trust between Users
• 3.2. Indirect Trust between Users
• 3.3. Preference Trust between Users
• 3.4 Users’ Direct Trust to Service
• 4. The Similarity Calculation
• 4.1. Similarity Calculation between Services
• 4.2. Preference Similarity Calculation between Users
• 5. Indirect Trust of the Target User to the Recommendation Services
• 6. Personalized Recommendation Algorithm Trusted Service
• 7. Experiment and Analysis
• 7.1. The Users’ Trust
• 7.2. Precision
• 7.3. Recall
• 7.4. F1_Measure Index
• 7.3. Analysis of Experimental Results
• 8. Conclusion
• Acknowledgements
• References