Database Monitoring Document Analytical Based On UIMA

This papers based on the current development of unstructured data mining research, analysis and abroad today on the progress of mining unstructured data. As various forms of unstructured data is difficult to use a unified tool to extract, by means of some auxiliary equipment must be before they can understand its content, through a variety of current market comparison of unstructured data mining tools as well as kinds of algorithm analysis, we hope to develop a specialized tool for the analysis of unstructured data, with the help of this tool can extract unstructured data and analysis. As the extensive database applications, and thus select a specific area of this non-text data as input text, the final OK to DB2 database monitor log as a research object, the goal is to unstructured data in the database were extracted. It makes database administrator to facilitate monitoring and management to ensure the efficient functioning of the database.Innovation of this paper is to propose a study to monitor the log database, combining popular framework for unstructured text analysis works, the method by combination of custom tags XML technology to complete and accurate expression to the unstructured data the implicit content of the framework in the Eclipse down to achieve a particular field based on unstructured data mining platform, which can be timely monitored from a large complex database administrators need to log in to obtain monitoring information to enable management Members of the problems that will be timely and appropriate treatment, to ensure that the database can be highly efficient and stable operation. Traditional unstructured data mining tool for the database area of the platform compared to the analysis, practical, strong, accurate data extraction for the text, and the platform has a high stability and security, which is extremely data management for database important to extend the platform is better, just add a custom tag can be independent, user-friendly.Platform is designed to achieve, the main focus of this paper is the development of unstructured IBM UIMA-based management framework, combined with text mining technology nowadays, by the custom tag database monitoring log to achieve the extraction of unstructured data, Combined with the method of XML database log data in the unstructured nature of the rules described in the definition is intended to enable the computer to recognize the paper we define, for the definition of the characteristics of unstructured data, this selection is the most simple and intuitive way, Which uses regular expressions, said the data, and finally through the integration of JAVA code, the entire platform design and implementation. Finally, the platform through the analysis of test results, we can see in this article the basic platform is designed to meet the initial design of unstructured data extraction purposes. Its main advantage is that it can quickly and efficiently extract data needed for the administrator, run faster, good stability, and the data for the field of professional, easy to use, less impact on system operation. Of course, there is still a lack of platform, there are many objective factors will affect the accuracy of analytical results, also need further details of the improvement. However, the overall functioning of the work from the platform point of view of non-structure for the future of data mining is a certain reference value.