Research And Application Of Fuzzy Neural Network In Student Performance Prediction

With the development of computer technology,the educational and teaching work in colleges and universities is increasingly dependent on the educational management system.While the educational management system stores a very large number of performance datas of Student course,how to effectively analyze the data to improve the quality of teaching in the school is a question which is worthy for deeply studying.In this thesis,the fuzzy neural network method is used to predict the students' corresponding graduation scores based on the usual grades,and the MATLAB software is used to simulate the predicted models.Finally,a set of perfect student achievement forecast management is developed based on this model which can achieve the purpose of student academic warning.Although fuzzy neural networks have two excellent functions of adaptive and nonlinear approximation which can solve the complex nonlinear relationship between course scores and graduation scores.However,fuzzy neural networks still have their own inevitable shortcomings in some aspects,such as redundancy between student score sample datas,low model operation effciency caoused by the high input varibale dimension.What'more,Fuzzy neural network is easy to fall into Locally optimal and difficult to find the global optimal value phenomenon.In this thesis,a fuzzy neural network model based on principal component analysis and genetic algorithm optimization is proposed for the above three problems.In the first step,the main component factor analysis method is used to reduce the dimensionality of the student's scores,so that the original high-dimensional data set becomes a low-dimensional new data set,thereby reducing the dimension of the model input variables and improving the efficiency of the model.At the same time,the process also solves the redundancy problem between sample data.In the second step,the genetic algorithm is used to optimize the parameters of the fuzzy neural network,so that the network can achieve global optimization.Thus,this thesis constructs an optimized fuzzy neural network performance prediction model.Then,the simulation experiment is carried out by MATLAB software.The results show that the model can make accurate predictions on the unknown graduation scores by using the existing student curriculum scores,so that the prediction results can be used as the basis to provide early warning for the students' academic performance.Thus the feasibility of the model for graduation prediction is verified.Finally,based on the optimization model constructed in this study,a set of student performance management system with high practicability,reusability and scalability is developed,So the students will be prompted for academic warning based on the prediction results.