基于Choquet积分的深层网数据源选择

摘要

This paper proposes a kind of data source selection strategies for Deep Web data integration based on Choquet integral. The quality of the Deep Web data sources is evaluated by the number of the query results, which is about to a set of keywords. This method of evaluation not only considers the weights between keywords, but also takes full account of the positive interaction which exists among multiple keywords. This way of the assessment of the Deep Web data sources is closer to the truth. Experimental results indicate that this method makes better performance than existing methods.%提出一种基于Choquet模糊积分的深层网数据源选择方法.通过在1组关键字上的查询结果评价深层网数据源的质量,评价时不仅考虑到关键字间的权重,同时还考虑多个关键字之间存在的交互作用.实验结果表明,该方法有效,且对深层网数据源的评价更接近真实.