基于网上特定话题的多侧面跟踪技术及应用研究

Research on Drift Analysis and Tracking Technique of Topic on Public Military Opinion

扩展对新媒体信息资源的获取和研究是数字图书馆系统进一步提升服务方式和质量的重要课题,而话题跟踪是其中一项关键技术。针对专门话题跟踪过程中的话题漂移及跟踪结果的杂乱无序问题,提出网络文本的客观中心和主观中心概念,给出话题主观中心的表示和更新方法,建立多中心模型按话题内容进行组织分析;提出先分类后聚类,依据客观中心跟踪话题和主观中心组织话题的方法对策,对传统KNN算法进行改进,把基于时间、地点、机构和关键词要素的KNN话题跟踪算法和基于内容要素的文本聚类过程进行合并,给出军事类话题多侧面跟踪组织的技术解决方案和实例。

Extending the acquisition and research of new media information resources is an important subject of digital library system. It is to further improve the service mode and quality. The topic tracking technology is one of the key technologies. The objective and subjective text center network center concept are put forward to solve the problem of topic drift and special topic tracking results in the tracking process. We have given the topic representation and subjective center update method,and establish a multi center model according to the topic of organizational analysis. The first classification after clustering, countermeasures according to the method of topic tracking and objective center the subjective topic center organization are put forward. We improve the traditional algorithm of KNN with time, location, and topic keywords elements. The military side tracking technology solutions have been studied.