论文部分内容阅读
在文本过滤中信息分流是提高过滤效率的强有力的手段 ,为此 ,提出了一种新的中文文本过滤的信息分流机制 .其基本思想是在概念扩充基础上 ,将不同用户的信息需求组织为树状结构 ,使其共同的部分成为共享分支 ,依据提出的侧面相似度和侧面匹配率来实现文本与模板的定量匹配 ,减弱传统的布尔模型对文本与模板匹配的严格限制 ,也弥补向量空间模型单纯数量化的不足 ,更加全面地反映用户的信息需求 .试验表明该机制能够明显地提高过滤效率 .
In the text filtering, information diversion is a powerful means to improve the filtering efficiency.Therefore, this paper proposes a new information filtering mechanism for Chinese text filtering.The basic idea is based on the concept of expansion, the information needs of different users organizations Is a tree structure, so that the common part of the tree becomes a shared branch. Based on the proposed side similarity and side matching rate, quantitative matching of the text and the template is achieved, thereby weakening the strict limitation of the traditional Boolean model on text and template matching and making up for the vector Spatial model simply quantify the lack of more fully reflect the user’s information needs.Experiments show that the mechanism can significantly improve the filtration efficiency.