社会化标签系统中基于本体的个性化推荐方法研究

Research on Personalized Recommendation Method Based on Ontology of Social

【目的】捕捉用户兴趣的动态性变化,优化个性化信息推荐效果。【应用背景】高效的个性化信息推荐方法可以根据用户行为特征主动为用户提供合适的信息资源,使信息的获取和利用更加快捷、准确。【方法】以"新浪微博"为例,通过挖掘用户及其关注者的微博数据,提取标签,计算二者兴趣相似度及亲密度,确定用户兴趣标签并优化标签描述,从而构建用户个性化"轻量级"本体,使得语义网资源能够准确地投放到用户界面。【结果】有效缓解了信息爆炸式增长所造成的"信息迷航"现象。【局限】微博数据中的杂音(广告转发、多语言描述)、数据不充分等,可能影响标签提取的准确性。

Objective] to capture the dynamic change of the user interest, to optimize the effect of personalized information recommendation. [background] Application of personalized information recommendation based on user behavior characteristics of the active method can provide the right information resources for users, so the acquisition and use of information is more efifcient, accurate. [method] Taking ＆quot;Sina micro-blog＆quot; as an example, through the data mining of micro-blog, users and their followers label extraction, calculation of two interest similarity and intimacy, determine the user label and optimize the description tag, so as to construct a personalized ＆quot;lightweight＆quot; ontology, the semantic web resource can accurately put into the user interface. [results] to effectively alleviate the explosive growth of information caused by the phenomenon of ＆quot;information＆quot;. [Limited] micro-blog data noise （AD forwarding, multi language）, the lack of data, may affect the accuracy of label extraction.