Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/77418| Title: | 利用使用者行為推論多個擴散網絡 Inferring Multiple Diffusion Networks with User Behavior |
| Authors: | Jui-Hsiu Hsu 許睿修 |
| Advisor: | 陳銘憲(Ming-Syan Chen) |
| Co-Advisor: | 王釧茹(Chuan-Ju Wang) |
| Keyword: | 社會網絡,網絡推論,擴散網絡,存活分析,使用者行為, social network,network inference,diffusion network,survival analysis,user behavior, |
| Publication Year : | 2021 |
| Degree: | 碩士 |
| Abstract: | 在社會網絡的領域內,有關網絡推論的問題,主要在於研究使用者間傳播資訊的影響力強弱。由於資訊擴散的軌跡是隱藏的,且不同主題間的資訊各有不同的傳播模式,所以需要透過大量的傳播資料來推論擴散網絡。而在過去的一些文獻內,已有提出許多方法來推論多個特定主題的擴散網絡。然而,這些方法僅考慮了資訊的擴散特性或使用者的屬性,卻未顧及使用者的資訊散播行為本身,是否會對資訊擴散造成影響。本篇論文,我們提出一個生成混合模型,名叫「UBRate」。它充分考慮了不同使用者間的行為差異和影響力來推論多個擴散網絡,且我們分別利用模擬資料和真實資料來驗證UBRate 的效益。最終,實驗結果成功表明UBRate 能有效地推論多個擴散網絡,並且優於其他基準模型。 The network inference problem (NIP) has been studied to infer hidden information diffusion networks. Since information diffusion patterns differ across topics, several approaches have been proposed to infer multiple topic-specific diffusion networks. However, these approaches consider only the diffusion properties of information or user attributes and ignore differences in user behavior. In this work, we propose UBRate, a generative mixture model which leverages the distribution and impact of distinct types of user behavior to infer multiple diffusion networks. Experimental results on both synthetic data and real-world data show that the proposed model effectively infers the diffusion networks and outperforms other baseline models. |
| URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/77418 |
| DOI: | 10.6342/NTU202100037 |
| Fulltext Rights: | 未授權 |
| Appears in Collections: | 資料科學學位學程 |
Files in This Item:
| File | Size | Format | |
|---|---|---|---|
| U0001-0901202116523200.pdf Restricted Access | 1.78 MB | Adobe PDF |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
