學術產出-Proceedings

Article View/Open

Publication Export

Google ScholarTM

政大圖書館

Citation Infomation

題名 Discovering Leaders from Social Network by Action Cascade
作者 Tsai, Ming-Feng;Tzeng, Chih-Wei;Chen, Arbee L.P.
蔡銘峰;曾智煒;陳良弼
貢獻者 資科系
日期 2012
上傳時間 22-Jun-2016 17:14:27 (UTC+8)
摘要 This paper proposes an approach to discovering community leaders in social network by means of a probabilistic time-based graph propagation model. To conduct the approach, we define an exponential decay function for influence as a function of time, and build action-specific influence chains by multiplying path propagation values; then, we create general influence chains by normalizing over all possible actions. In the study, our approach identifies community leaders as those people whose initiated influence chains are relatively more than the chains they involved. In our experiments, a small Facebook network dataset with 134 nodes and 517 edges is employed to assess the performance of the proposed method. In addition, several baselines are also carried out for comparison, including three naive and one user-involved approaches. The experimental results show that, compared with the baselines, the proposed method can effectively identify community leaders within the social network, achieving 0.8 in terms of F-measure.
關聯 Proceedings of the Fifth Workshop on Social Network Systems (SNS `12), 12:1-12:2, 2012
資料類型 conference
DOI http://dx.doi.org/10.1145/2181176.2181188
dc.contributor 資科系
dc.creator (作者) Tsai, Ming-Feng;Tzeng, Chih-Wei;Chen, Arbee L.P.
dc.creator (作者) 蔡銘峰;曾智煒;陳良弼zh_TW
dc.date (日期) 2012
dc.date.accessioned 22-Jun-2016 17:14:27 (UTC+8)-
dc.date.available 22-Jun-2016 17:14:27 (UTC+8)-
dc.date.issued (上傳時間) 22-Jun-2016 17:14:27 (UTC+8)-
dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/98234-
dc.description.abstract (摘要) This paper proposes an approach to discovering community leaders in social network by means of a probabilistic time-based graph propagation model. To conduct the approach, we define an exponential decay function for influence as a function of time, and build action-specific influence chains by multiplying path propagation values; then, we create general influence chains by normalizing over all possible actions. In the study, our approach identifies community leaders as those people whose initiated influence chains are relatively more than the chains they involved. In our experiments, a small Facebook network dataset with 134 nodes and 517 edges is employed to assess the performance of the proposed method. In addition, several baselines are also carried out for comparison, including three naive and one user-involved approaches. The experimental results show that, compared with the baselines, the proposed method can effectively identify community leaders within the social network, achieving 0.8 in terms of F-measure.
dc.format.extent 615036 bytes-
dc.format.mimetype application/pdf-
dc.relation (關聯) Proceedings of the Fifth Workshop on Social Network Systems (SNS `12), 12:1-12:2, 2012
dc.title (題名) Discovering Leaders from Social Network by Action Cascade
dc.type (資料類型) conference
dc.identifier.doi (DOI) 10.1145/2181176.2181188
dc.doi.uri (DOI) http://dx.doi.org/10.1145/2181176.2181188