摘要
针对现有中文Deep Web查询接口的模式匹配方法准确度不高、效率较低、自动化不够等问题,提出了一种基于AHPH的中文Deep Web模式匹配方法。该方法通过对属性进行配对后计算各个属性匹配对的相似度,根据一定的规则获取最优匹配。针对属性配对的相似度计算,采用基于《知网》(Hownet)的词语相似度计算方法得到属性词语之间的各个相似度,并利用层次分析法(AHP)为属性词汇之间的各个相似度分配权重。实验结果表明,该方法能明显提高模式匹配的精确度和召回率,有效地提高了匹配质量。
in order to solve the low accurate and efficiency in Chinese Deep Web query Interface matching, a approach of Deep Web query Interfaces matching based on AHPH is proposed. This method calculates similarity of paired attributes defined, and gets the best mateh According to certain rules. To calculate the similarity of paired attributes defined, the similarities of the words of the attributes is obtained by using a word similarity algorithm based on Hownet, and weights for these similarities are distributed using AHP. Experimental results show that the approach is highly accurate and effective.
出处
《计算机工程与设计》
CSCD
北大核心
2013年第1期293-297,共5页
Computer Engineering and Design
基金
国家科技重大专项课题基金项目(2009ZX03001-019-01)