基于SimHash和混合相似度的多模式匹配方法

Multiple schema matching method based on Sim Hash and mixed similarity

为了解决多源异构民航旅客服务数据集成过程中存在多模式匹配的效率不高、精确性不足、完整模式信息获取难度较大等问题,提出了一种基于Sim Hash和混合相似度的多模式匹配方法。该方法首先基于PMI计算特征单元权重,并通过Sim Hash算法构造属性列的签名来表示属性特征,以降低特征维度,进而引入K-means++算法对属性聚类并生成候选匹配集。最后基于属性的混合相似度构建属性映射图,以直观的方式展示属性间的匹配关系,同时提高多模式匹配效率。实验结果表明该方法具有可行性,为高效地解决多源异构民航旅客服务数据集成中的模式冲突问题提供新的解决方案。

In order to solve the problems of multiple schema matching in the process of integrating multi-source heterogeneous civil aviation passenger service data,such as low efficiency,low accuracy and the complexity of obtaining complete schema information,this paper proposed the multiple schema matching method based on SimHash and mixed similarity.Firstly,the method calculated the weight of feature units based on PMI,and generated the signature of columns by SimHash to represent attribute features to reduce feature dimension.Further,it employed K-means++to generate candidate matching sets by clustering the columns.Finally,it constructed the mapping graph of attributes based on attributes’mixed similarity,and displayed the matching relationship between attributes intuitively.Meanwhile,it improved efficiency of multiple schema matching.The experimental results verify the feasibility of the proposed method.The method provides a new solution for efficiently resolving the schema conflict in the process of integrating multi-source heterogeneous civil aviation passenger service data.