Application On Identification Of Poor Students Based On Rough Set And BP Neural Network

Our government has been dedicating to improve the funding policy system for needystudents since1987. From2007, the government put more efforts to increase the subsidyamount and the proportion of funding of state grants for needy student. At present, thetraditional way to identity college students with financial difficulties has many disadvantages,such as subjective judgment, human influence factors, and the lack of scientificity andoperability. Therefore, it is very important and urgent to create a fair and effective way toidentify the students who are real in need.In recent years, researchers in this area use attribute recognition theory, gray relationalanalysis, the multi-index comprehensive evaluation of fuzzy mathematics, step analysis,decision tree classifier, logistic regression model, BP neural network model to identify thepoor college students in China. Among these theories and models, researchers are moreinterested in BP neural network model because it needs no assumptions at all. However, whenthe scale of neural network is large and there are more training sample set, the training timewill be lengthened, and the classification speed will be slow. Based on the literature review,this paper tries to promote a new method derived from the neural network based on rough setmodel to solve the mentioned problem, which, with the classification ability unchanged,employs the rough set theory to achieve attribute reduction of the sample set, to reduce thedimension of the sample, and to simplify the neural network input nodes.Firstly, the author studies the data of needy students of a particular college and analyzesthe attribute information of data recorded in the needy students. Secondly, the author tries tobuild up the data pool with MS Office Excel. Then use the Rosetta to make attribute reductionof the decision table, study and predict the experimental data source as the next step. At last,by constructing BP neural network, the paper create a nonlinear mapping between theeconomic conditions of college students and the needy students identifying. This paper tries toprovide a better way to identify college students with economic difficulties by studying andrealizing the model.