Conference proceeding
Information passing in online recommendation
Proceedings of the 1st workshop on user engagement optimization, pp 3-6
01 Nov 2013
Abstract
In this paper, we analyze a recommendation network with over 4,000 users and half a million books. There are two types of edges in this network, "read" relations between users and books, and following relations between users. We first investigate in general, if one's followees' recommendations have impacts on one's decision. We then analyze the correlation between one's influence and her centrality in the network. Finally, we study how effective a recommendation is as one sends or receives more and more recommendations. Results show that although in general, one's followee do have an impact over her decision, such influence is not correlated with the followee's centrality. As one receives more and more recommendations for a product, it is more likely that she will accept it. However, there is a saturate point over which more recommendations will have no further impact. As one sends out more and more recommendations, the probabilities that these recommendations get accepted become larger and larger.
Metrics
10 Record Views
2 citations in Scopus
Details
- Title
- Information passing in online recommendation
- Creators
- Jia Huang - Drexel UniversityXiaohua Hu - Drexel University
- Publication Details
- Proceedings of the 1st workshop on user engagement optimization, pp 3-6
- Conference
- 1st workshop on user engagement optimization, 1st
- Series
- UEO '13
- Publisher
- Association for Computing Machinery (ACM)
- Number of pages
- 1
- Resource Type
- Conference proceeding
- Language
- English
- Academic Unit
- Information Science
- Scopus ID
- 2-s2.0-84889579467
- Other Identifier
- 991019174763204721