Research Of Pattern Extraction From Semi-structured Data Based On Rules

With the continuous development of information technology in recent years, the form of data such as semi-structured data has emerged. Non-pattern and self-description have made semi-structured data easy to use, meanwhile, they also bring difficulties in integration of structured data and semi-structured data. However, the key to solve the data integration problem is that pattern extraction of semi-structured data. Therefore, how to extract pattern from semi-structured data accurately has become one of the hot research.In response to these problems, with the report as an example, this paper propose a method of pattern extraction of semi-structured data based on rules, defines a generic report description language to solve the problem of pattern extraction of semi-structured data. We carry out a study on the theory of report description method, the representation of rules, the storage structure of rules, conflict management of rules. Specific studies are as follows:1. Defines a generic report description language. The language combines the most advanced technology of forms and the actual situation of oil exploration work in a large number of complex business logic, with the features of advanced, practical, scalability, etc. It is the result of research of a large number of forms of actual business needs on the whole oil exploration industry and through the test of the practical application to prove its usefulness.2. Through analysis of relevant theories of rules, propose our own theory of report description method, the representation of rules, the storage structure of rules, conflict management of rules. Use the rules to describe the model of semi-structured data, through the interpretation of the rules eventually to generate the appropriate report description language and to realize pattern extraction of semi-structured data.3. Introduce data element dictionary and data dictionary to describe the business logic. Through the establishment of common data elements dictionary and data dictionary to interpret the information of the form, using them to complete the description of business logic of the form.Finally, based on the research of the paper synthetically, it presents the achievement condition of the system. It verifies the feasibility and effectiveness of the method which is proposed in this paper.