文献类型：期刊文献

英文题名：Research on Key Technology of Web Hierarchical Topic Detection and Evolution Based on Behaviour Tracking Analysis

作者：Chen, M.[1]

第一作者：陈默

通讯作者：Chen, M[1]

机构：[1]Beijing Union Univ, Business Coll, A3 Yanjingdongli, Beijing 100025, Peoples R China

第一机构：北京联合大学商务学院

通讯机构：[1]corresponding author), Beijing Union Univ, Business Coll, A3 Yanjingdongli, Beijing 100025, Peoples R China.|[1141721]北京联合大学商务学院;[11417]北京联合大学;

年份：2019

卷号：14

期号：3

起止页码：311-328

外文期刊名：INTERNATIONAL JOURNAL OF COMPUTERS COMMUNICATIONS & CONTROL

收录：;Scopus(收录号：2-s2.0-85067017751);WOS:【SCI-EXPANDED(收录号:WOS:000469927900002)】；

基金：This paper is supported by National Natural Science Foundation of China under Grant Nos.71572015, Support Project of High-Level Teachers in Beijing Municipal Universities in the Period of 13th Five-Year Plan under Grant Nos.CIT&TCD201704072, Premium Funding Project for Academic Human Resources Development in Beijing Union University under Grant Nos.BPHR2018AS01.

语种：英文

外文关键词：Web hierarchical topic; topic detection; event evolution; behaviour tracking analysis

摘要：In the development background of today's big data era, the research direction of Web hierarchical topic detection and evolution characterized by the semi-structured or unstructured data has caught wide attention for academicians. This paper proposes an idea of Web hierarchical topic detection and evolution based on behaviour tracking analysis taking the network big data as the research object, and expounds main implementation methods, which include the instance analysis of the usage mode, the instance analysis of the seed, the set analysis of similar instance supporting the topics, the set analysis of similar instance supporting the events, the evolution analysis of the event, and expounds the algorithm of Web hierarchical topic detection and evolution based on behaviour tracking analysis. The process of experimental analysis is organized as follows, first of all, the experiment analyses the quality of topic detection, the accuracy rate with the number of instance concerned and the seed threshold variation trend, the accuracy rate with the number of instance concerned and the probability threshold variation trend, secondly, the experiment analyses the quality of topic evolution, the accuracy rate with the variation trend of parameter adjustment, the accuracy rate with the number of instance concerned and the similar threshold variation trend, finally, the experiment analyses the time consuming to solve main research problem under different method, the qualitative result of topic detection and evolution under different data set. The results of experimental analysis show the idea is feasible, verifiable and superior, which plays a major role in reconfiguring Web hierarchical topic corpus and providing an intelligent big data warehouse for the network information evolution application.