A fusion of predicate logic and document semantic distance method orientated on data and context mapping

With the continuous development of network and database technology, share and conversion on heterogeneous data are still a complex problem currently facing. Current methods are mostly based on the similarity between the field to complete the matching process, although this method can finish partial matches on information, its time complexity is very high due to the large processing data sets and the lack of prior knowledges in exploitation degree. Therefore, the information matching method is concentrated on this paper on the basis of neural network. The method adopts not only semantic similarity and first-order logic in the process of table match, but also SOM+ to produce feedback neural network in order to realize field match. The method can efficiently decrease the time complexity. At the same time, training time of neural network can be reduced with the help of history match data. At the same time the accuracy of matching is also improved.