Preference Data Analysis And Mining On National College Entrance Examination And Admission

How to decide the preference to increase the probability of dream college matriculation is concerned by examinees and their families closely. Although there are corresponding auxiliary references systems for students'preference decision in National College Entrance Examination and Admission (NCEEA).They just make a comparison of the historical score or precedence data so simple that they can't support help for students fundamentally. In view of this, by employing data warehouse, OLAP and data mining technologies on the historical NCEEA data which include plenty of student's personal information, preference data, major information and matriculation etc., some interesting and referential results are mined.The contributions of this paper are in the following four aspects:1.Cleaning the historical data in NCEEA systemA comparison of recent nine years historical enrolling data has been made to find the some problems in data quality and integration; the data code tables in national college entrance examination and enrollment system has been standardized based on the advice of professional staff; Finally, historical data have been extracted, transformed cleaned with SSIS (Microsoft SQL Server 2008 Integration Services) according to the standard code tables then loaded to data warehouse.2. Establishing the entrance examination data warehouseThis research subject which is a part of the key technologies research and development program of Chongqing comes from a practical application project; therefore requirement has been analyzed scientifically; the subjects has been made sure according to users'requirement; the granulation, measurements and the data warehouse model have been designed. The cleaned data have been loaded to data warehouse finally.3.Conducting research on preference data analysis on NCEEA Multi-dimensional dataset has been modeled according to the preference analysis requirement. Preference cube has been established after loading the data from the data warehouse to multi-dimensional model. By employing the OLAP technology on the historical preference data, some interesting and referential results are mined based on the analysis of college application heat degree and minimum score of college matriculation.4. Conducting research on preference data mining on NCEEAPreference data mining model has been established according to Cross-industry Standard Process for Data Mining (CRISP-DM).By employing the Microsoft decision trees algorithm on the preference data and college matriculation results in NCEEA, matriculation preference and key factors which have influent on college matriculation are mined. The experiment on preference data mining model demonstrates that our data mining model has better predicting accuracy by comparing with naive Bayes classification and neural network classification. It also shows that some interesting and referential results are mined.