摘要:
Web document clustering is one of the most important research branches of Clustering Analyzing. The objective of web document clustering is to meet the need of retrieving web document efficiently from massive information in Internet. Recently social tagging is the important form of document organization in web 2.0, and the tagging as a document descriptor is used to improve the effectiveness of web searching. But a web document usually belongs to various category of tagging, which may lead to the difficulty of browsing web document based on single tagging. This paper explores the use of Formal Concept Analysis (FCA) as mathematical tool to analyze the social tagging of web document, and presents a model for web document clustering based on tagging semantic. Furthermore, taking community web site Douban as an example, the model is applied to allow users to tag and serendipitously browse web document using Formal Concept Analysis.
期刊:
Lecture Notes in Electrical Engineering,2012年124 LNEE(VOL. 1):383-388 ISSN:1876-1100
作者机构:
[Li, Xiao-Yun; Yu, Ying; Ma, Jia-Yu; Yang, Xiao-Hua; Wan, Ya-Ping; Liu, Zhi-Ming; Jiang, Hui] School of Computer Science and Technology, University of South China, Hengyang, Hunan, China
作者:
Zhiming Liu 0003;Yaping Wan;Jichang Sha;Xiaohua Yang
期刊:
International Journal of Information Processing and Management,2010年1(2):58-67 ISSN:2093-4009
通讯作者:
Wan, Y.
作者机构:
College of Information System and Management, National University of Defense Technology, Changsha 410073, China;Computer Department, HuaZhong University of Science and Technology, Wuhan, 430074, China;School of Computer Science and Technology, Nanhua University, Hengyang 421001, China
通讯机构:
School of Computer Science and Technology, Nanhua University, China
摘要:
Peer-to-Peer(P2P) systems have been one of the hottest research issues for its attractive advantages, such as self-organizing, scalability and fault tolerance. However, super node failure is a serious problem because of the high dynamic and untrusted peers. A super node election mechanism is implemented, which included four basic components: The registration center module, group membership maintenance module, fault detector and super node election algorithm. The system constantly detects failure, at the same time, gathers the statistics of the online rate and fault detection message response time of each node. It only takes less computing and storage resource to master the node "active" state and distribution, without increasing too much additional works. The results show that average service recovery time and average error rates of a super node election service presented by this paper are less than Toung's S1, while the availability of service is close or slightly lower than the S3. It achieves the desired effect of the experiment that while maintaining the availability under certain conditions, it can reduce error rate and service recovery time, thereby increasing the reliability of P2P system. In addition, this realization of the super node service mechanism can be deployed flexibly according to user needs and can provide convenience for the future implementation of solutions.