A Scalable Matching Approach Based Density Function for Heterogeneous Database Integration

Scalable matching is used for summarizing information in large databases such as data warehousing and transaction management. In this paper, we propose a matching approach for integrating database models. We adopt the mixture density function, which basically addresses the number of general properties that attribute matching should fulfill. This approach of matching refers a scenario to identify the common portion between the source databases and perform the queries over original databases. In our approach, we first construct a model with shared ontology by analyzing the matching instance between attributes. We finally present how this approach can be actually used to perform operations over distributed databases. We also exemplify this approach with local maxima and probabilities of density estimation of EM (Expectation - Maximization) algorithm.