Research And Implementation Of Decision Support System In College Discipline Construction Based On Data Mining

Disciplinary is the basic unit of college’s learning organizations. And it also is theintegration point of personnel training and the scientific cultural development.Strengthening the disciplines has become the development trend of the contemporaryworld of higher education. With the continuous progress of computer technology,network technology and the information age, the institutions of higher learning havebeen achieved information management. The use of a variety of managementinformation system composed of a huge factory which could produce a large numberof effective analysis data. How to use the continuously accumulating informationservices for discipline construction of the colleges and universities and improve theobjective and scientific decision-making of the university discipline management, ithas become a problems to be solved by the college disciplinary managers and has alsobecome a hot topic of research scholars.With the college disciplinary construction as application background, This papermainly research how to effectively use the data generated by the universitymanagement information system to provide scientific decision-making for theuniversity management. The topic in-depthly researched data mining technology anddecision support systems technology to explore the trends and phenomenon of thesubject development, reveal rules and characters of the development of highereducation，and inherit the new development of decision support systems in the theuniversity discipline management.In this paper, the decision support system based on data warehouse is applied upto the construction of university. And the system architecture of the collegedisciplinary construction decision support system was completed. The systemarchitecture not only followed the traditional architecture of architecture but alsoexpanded it, and it effectively combined with the data mining technology and webtechnology. The system of this paper not only can provide effective decision supportcapabilities but also has a good scalability and the interaction ability. Secondly, thesubject used the historical data provided by the Beijing key disciplines platform andthe Beijing Master and Dr. Platform as a data source, studied data warehousemultidimensional data model of university discipline construction, established eightfact tables, completed data warehouse of the college discipline construction. Afterbuilded the data warehouse, the topic research data mining decision tree algorithm,applied the bayesian theory to prune the decision tree created by decision tree C4.5algorithm, and provided a decision tree pruning algorithm based on bayesian theory.The method used the Bayesian posterior theorem to deal with the each branch of thedecision tree defines for two types of validation, and cut off branches disatisfied withthe condition from the decision tree so that the decision tree get leaner. The subjecthas established the prediction model key disciplines approval using the algorithm andhas been validated the validity prediction model by experiments. After gained the prediction model, the subject studied multistage fuzzy evaluation method based onrough set to to establish quality comprehensive evaluation model of universitydisciplines according to the discipline integrated evaluation index system promulgatedby the City Municipal Education Commission. The evaluation model can make acomprehensive evaluation of the discipline and can be found further comprehensiveranking.Finally, combined with the decision support system architecture of collegedisciplinary construction, the data warehouse of college disciplinary construction andthe data mining algorithms, based on the J2EE standard enterprise system standard,used the MVC pattern, this topic completed and achieved a college disciplineconstruction of decision support system based on B/S structure.