Information retrieval for probabilistic pattern matching based on neural network

Distributed information systems tend to be highly heterogeneous, integrating different computer platforms, data storage structure, different database management systems and schemas which converted data structure under one model to data structure under a different model. For these reasons, a probabilistic pattern matching framework based on neural network is introduced. Our approach gives a probabilistic interpretation of the prediction weights of the candidates, selects the rule set with highest matching probability. Pattern matching is the problem of finding correspondences (mapping rules, e.g. logical formulae) between heterogeneous schemas e.g. in the data exchange domain. Through the union formulae, distributed information system may achieve uniform retrieval and provide a strong reference for the research of information retrieval.