摘要
针对Web2.0社会化标签系统中标签组织混乱和标签语义模糊的问题,围绕标签共现网络的拓扑图,建立了一种形式化的标签语义相关度计算模型.该模型利用基于统计的标签语义相关度计算结果,为标签共现网络的拓扑图扩展以语义相关度权值,并定义算子围绕标签共现网络的拓扑图来计算权值的综合效应,从而可以显式地描述标签语义关联的交叉影响,并在标签语义相关度的计算中融入这些影响因素.从照片共享网站Flickr中抓取热门标签数据,通过实验对该模型的计算过程、计算结果的有效性和实用性等进行了分析评价.实验表明,该模型的标签语义相关度计算结果更为准确,可以更好地引导和约束Web2.0用户的标签使用行为.
Regarding the problems of the structureless organization and implicit meaning of tags in the social tagging systems in Web2.0, a topological graph based formal model of semantic relatedness measure is proposed to fully exploit the interplay of the semantic co-relations among a large number of tags. In this model, the results of the statistics based tag relatedness measures are used to extend the topological graph of tags co-occurrence network with weights of edges, two operators are invented to synthetically compute the overall effect of the weights within the extended graph, so that the interplay of semantic co-relations of tags can be explicitly represented and the semantic relatedness of tags can also be measured soundly. To illustrate the calculating process and to testify the validness and feasibility of the calculating results for this model, an experiment is conducted with the set of the most popular tags crawled from Flickr. com, a famous photos Sharing website. Experimental results show that the model can lead to better results, and is highly applicable to the guidance and constraint of annotating behaviors in Web2.0 environments.
出处
《西安电子科技大学学报》
EI
CAS
CSCD
北大核心
2012年第3期196-201,共6页
Journal of Xidian University
基金
国家自然科学基金资助项目(61101143/F010202)
中央高校基本科研业务费专项资金资助项目(CHD2012JC022)
