Design Of College Financial Decision Software Based On Data Mining

With great emphasis from the government and the community, the university managementshould advance with the times. Colleges and universities must conduct a scientific,standardized reform in the aspects as idea, target and methods, student payment system, poorstudent information systems and other database provide data basis for data mining, whichgreatly enhance the college financial data processing capability, and the financial decisioncould be implemented.This article analyzes student payment behavior based on in-depth analysis and researcheson data mining. With data mining algorithms, we make arrears amounts clustering byanalyzing payments of premiums, and predict the proporting of colleges or professional, andfinally discover useful information from the background database through graphic display,this could provide reference basis for managers and policy makers to analyze.In this article, with student fee system and poor student information system as thesimulation training sample sets, we use DBSCAN algorithm for clustering analysis, the ID3algorithm to forecast and analyze, and JFreeChart algorithm for clustering analysis, the ID3algorithm to forecast and analyze, and JFreeChart to design data mining graphical display,through database and Java programming language to implement the extract, conversion, query,forecasting, analysis and presentation. The research and design in the university financialdecision support system is good valuable.Firstly, build the collection of data sources, design data analysis model, note that datasources are inconsistent, the database table structure is inconsistent, and the existence ofredundant data on the database, we put forward a problem which is the most attractive forcolleges as the test point, in this article we choose payment of students and poor studentsinformation system for the test object, in order to test the feasibility of the entire system.Secondly, Density-Based Spatial Clusering of Applications with Noise(DBSCAN)algorithm is density-based clustering algorithm. With this algorithm, take the credit amount asthe radius threshold to cluster the test samples of the different situation of arrears, and makeappropriate improvements to the sensitive issues of the density threshold for the algorithm,the clustering results can be used as a basis for analysis of the credit amount.Thirdly, ID3algorithm is Decision Tree Algorithm, we need generate decision trees forma data set to get an effective classifier, by entering the appropriate data sets, data objectscommon value, and object eigenvalue forecase analysis is executable. Finally, we can realize university financial decision support system on the base of thefront support, to display decision support system; we can display the process of data miningof payment of students through pie chart or bar chart.