Based On Neural Networks Of Heterogeneous Database Semantic Discovery And Matching Technology

Applications in a wide variety of industries require access to multiple heterogeneous distributed databases. During the process of constructing integration system for enterprise heterogeneous databases, Semantic integration is necessary to guarantee the validity and integrality of integrated data. But now it faces two main issues:One is how to identify corresponding attributes in different databases that represent the same real world concept.The other is how to improve the automated degree.Based on the above two points, we have applied artificial neural network (Ann) to deal with the above two main problems. Using artificial neural network to learn from samples data, then we can get rules to identify corresponding attributes in different databases. The aim is to minimize the human participation during the integration process. Our research work focuses on the following four aspects:1) Semantic discovering and corresponding methods for heterogeneous databaseReferring to the other research methods, we adopt SOM and BP neural network as the main body and XML as the description method for metadata.2) Semantic description methods based on metadataWe construct a semantic model of metadata to describe attributes characters. The model includes data schema and data content statistic.3) Semantic classification model based SOM networkWe use the classification model to combines attributes within a database. This is done using an unsupervised learning algorithm. The output is used as training data for the next stage.4) Semantic discovering and matching model based BP network The classifier output is used as training data for a BP neural net. The net produced by this can recognize attributes within the database based on their metadata and emerge learning rules. Then the rules are applied to identify corresponding attributes in different databases.