【摘要】 针对单纯的Pre-query和单纯的Post-query分类方法对中文Deep Web数据源自动分类的不足,本文提出了三段式逐步求精的中文Deep Web数据源自动分类方法,并且基于该方法设计了相应的分类系统。通过实验表明:对于实验数据源,这种自动分类方法比单纯的Pre-query分类方法、单纯的Post-query分类方法以及无影响度差别的文献[2]的分类方法的查全率、查准率和F1值都要高。更多还原

【Abstract】 This paper presents a three-phase gradual refining data source automatic classification of Chinese Deep Web,and designs corresponding system based on this method.It has improved simple Pre-query and simple Post-query classification method.Eperiments show that:this method has the highest recall rate,the highest precise rate and the highest F1 value,among simple Pre-query,simple Post-query,our method and document’s method that has no difference in effect on classification,for experiment data source.更多还原