VSM的权重改进对文档相似度的影响研究

向量空间模型是以索引项权重为核心的模型,索引项权重对文本分类、检索等的效果起着重要的作用。文中使用了一个基于关键词的权重,并利用它改进传统向量空间模型的权重算法。改进后的模型综合考虑原有索引项权重和文档中关键词的权重。在特定领域FAQ的检索中作测试实验,结果表明,改进的方法提高了检索的查准率、查全率。

Research of Documents Similarity Influence Based on Improved VSM Weight

The terms weight is the core in VSM ,it plays the important role in text classification,text retrieval,etc.A new weight based on key is put forward,so as to improve the weight formula of VSM.Further more,original characteristic terms weight is also combined in the new VSM.With the test based on special domain FAQ,Experiment results show that the improved method raised the precision,recall and the F test value.