Search Of Enterprise Data Based On Metadata

Computer technoloies have made an innovative development in the construction of enterprise information, have made possible to use information more easily and conveniently, and as a result, the amount of information start to increase at an explosive rate, which results in difficulties in using information. Although search engine is an effective way to address the problems, there still exist many problems, which can be catagoried into two class. First, due to the particuliarity, specialty and territoriality of the enterprise environment, the diversity of enterprise data ranges from documentation, database to email; heterogeneity of data, the growth of unstructured-data goes with the structured-data as well, and second, there still exist problems when search engine indentifying enterprise resource, understanding enterprise information. So in order to search enterprise data effectively, we have to address the problems above.In this thesis, we proposed a Meta-data-based Enterprise Search Engine. A Meta-data Web could be established by exploiting the management data, Meta-data, produced in the process of the construction of enterprise information. It consists of two different models of Meta-data, unified management model, web model and the transformation between them. The unified management model accomplishes the management of Meta-data from heterogeneous system; The web model implements the representation of the relations between Meta-data; And the transformation part achieves the transformation from unified model to web model. Both the representation and expression of the enterprise data can be achieved by using Meda-data web. That will make datas within some special fields easier to understand. Mean while, an algorithm based on Meta-data Web was also proposed in this thesis, which can improve the efficiency, recall and precision of searching.