Design And Implementation Of Data Mining System For Beijing Common High School Uniform Examination

Beijing Common High School Uniform Examination, being a criterion-referenced Large Scale test, aims to measure and value school work of high school students and quality of school teaching, and it is the basis used to weigh whether students reach high school graduation standards. With the extensive application of database system in the united examination management, Beijing education examine center accumulate large number of relevant data about the united examination. Because lack of effective processing means, these data become rare revisited data files after the end of the business activities. Data mining technology introduced in data analysis of the united examination, analysis of sedimentary data, find out hidden knowledge and potential rules about data, provides data support for scientific evaluation of the role of the united examination in senior high school teaching and for education administration departments making relative policy.The key of Data Mining is to identify a suitable goal of mining. In order to identify the goal of mining it is not only to consider business management needs, but also to consider the characteristics of data to ensure that the objectives data can be achieved on the basis of available. The main objective is to through analysis of different examination time (normal examination, non-normal examination), different types of candidates (Normal, private), different forms of examination (Examination, the group self-examination) of the test scores, to support the existing policy changes and to improve the provision of data, so that High School Certificate of Education Examination Of Beijing can really play high school examination of education quality monitoring and evaluation role.First of all, collect relevant data, which from the main business system of Education Examination, in order to meet the needs of mining theme may be required from other business systems access to data. The existing business system is mainly used relational data model, the need for excavation may be needed model conversion. Secondly, the need to clean up the original data, remove the abnormal data, such as null and unreasonable value; need for simplification of the original data, delete and not relevant to the subject of data mining.Identification of the target and pretreatment of the data mining, then below is a suitable choice to meet the needs of mining algorithm, on the base of researching several important characteristics of the mining algorithms, the application of algorithms, the algorithm to achieve, and so On. Two major methods: correlation analysis and classify analysis are studied. How effective use of mining algorithms to solve practical problems is the focus of the study.Finally, the results of mining is analyzed to determine the meaning of their representatives and determine whether it has a practical significance.