摘要
Abstract
With the continuous development of information technology,the prediction of college students' academic situation be-comes an important issue in the process of college education management and teaching.In order to solve the problems of single mode and insufficient depth of data in traditional academic situation prediction,the study constructs a prediction model of col-lege students' academic situation based on deep belief network(DBN).The proposed model is compared with k-means cluste-ring algorithm and neighboring classification algorithm.The real values are compared with the model predicted to verify the performance of the model.The results show that the detection rate,error rate,PR curve and F1 values of the proposed model are 96.53％,7.52％,0.81 and 86.53％,respectively,the value of the squared correlation coefficient R2 of the proposed model in the training set is 0.9638,and the value of the mean squared error(MSE)is 0.005 39,which indicates the high application value of the proposed model.The reliability and practical application performance of the proposed model in predicting the aca-demic situation of college students are verified.关键词
数据挖掘/教育管理/学情预测/深度置信网络算法Key words
data mining/education management/academic situation prediction/DBN algorithm分类
信息技术与安全科学
