论文部分内容阅读
信息过滤是卫生政策知识服务平台建设中的核心技术,在系统研究信息过滤的几种经典方法的基础上,确立将向量空间模型作为该平台的信息过滤方法,并进行一定的改进,以避免传统向量空间模型的不足。在字段间权重设定方面,采用信息检索过程中评价检索效果的两个经典指标,即查全率和查准率进行过滤效果的评价,并进行反复测试,最终确定各类资源不同字段在信息过滤过程中设置的权重及阈值,成功完成信息采集、信息分类、信息主动推送等功能。
Information filtering is the core technology in the construction of health policy knowledge service platform. Based on the systematic study of several classical methods of information filtering, the vector space model is established as the information filtering method of the platform and some improvements are made to avoid the traditional Insufficient vector space model. In terms of weight setting between fields, two classical indexes, that is recall rate and accuracy rate, are used to evaluate the filtering effect in the process of information retrieval and repeated testing is performed to determine the different fields of information The weights and thresholds set in the filtering process successfully complete the functions of collecting information, classifying information, and actively pushing information.