面对海量的网络新闻信息，为了能更加准确与全面地从中发现用户感兴趣的话题，提出一种基于事件关联网络的用户兴趣话题发现算法。该算法建立了代表事件之间关联关系的事件关联网络，基于该事件关联网络，采用链接分析技术度量用户对不同新闻事件感兴趣的程度，从而采用针对新闻特定语义架构的改进Single⁃pass聚类算法发现用户感兴趣的话题。此外，采用Bootstrapping算法，实现对相关兴趣领域词汇的语义扩展。实验表明，该算法能够更加准确而全面地获取用户感兴趣的话题。%Being faced of massive Internet news information,to improve the accuracy of detecting the topics that the users are interested,a topic detection algorithm based on the network associated with the events is proposed for users’interest. The algorithm established an event⁃related network representative of relevance relationship among news events. The link analysis tech⁃nique is used to measure the degree of user interest in the news,so as to identify the topics that the users are interested by using an improved Single⁃pass clustering algorithm based on news specific semantic structure. In addition,Bootstrapping algo⁃rithm is adopted to achieve the related interest words’semantic extensions. The experiment result shows that the algorithm can more accurately and comprehensively get the topics that the users are interested.
展开▼
