基于遗传神经网络的学生成绩预测

Student scores prediction based on genetic neural network

提出一种基于遗传神经网络的成绩预测方法。首先用相关分析法计算了基础课程成绩与目标课程成绩的相关系数,选取了与目标课程成绩相关度高的基础课程成绩作为输入项,然后引入遗传算法对反向传播(back propagation,BP)神经网络的初始权值和阈值进行优化,实现学生目标课程成绩预测。采集某大学通信工程专业1601—1603班级99名学生的实际教学数据进行验证,结果表明,该方法与BP神经网络模型相比,预测的均方根误差由8.6降低为2.8。

A scores prediction method based on genetic neural network is proposed. Firstly the correlation coefficient between the basic course score and the target course score is calculated by the correlation analysis method, then the basic course scores with high relevance to the target course score are selected as an input. Secondly, the genetic algorithm for the initial weights and thresholds of the (back propagation, BP) neural network are optimized to establish a genetic neural network student scores prediction model and student's target course score can then be predicted . The validity of the model was verified by the actual teaching data from 99 students in the 1601-1603 class of communication engineering module in one university. Compared with the BP neural network model, results show that the predicted root mean square error is reduced from 8.6 to 2.8 by this model.