基于熵权-离差最大化的BP神经网络编程能力评估方法
BP Neural Network Programming Capability Evaluation Method Based on Entropy Weight-deviation Maximization
许超焕 1许新华 1石沁语 1乔凯 1虞烨青1
作者信息
- 1. 湖北师范大学计算机与信息工程学院,黄石,435002
- 折叠
摘要
针对传统教学中编程能力难以评估的问题,本文提出一种使用反向传播(BP)神经网络评估学生编程能力的方法,组合熵权法和离差最大化法得出编程能力各指标权重并计算出学生编程能力综合得分,以优化评估模型的输入,通过调整隐含层个数增大模型输入空间,以减小评估模型输出误差.比较不同结构BP神经网络的评估效果,实验结果表明,隐含层个数为2时评估模型输出误差最小,模型预测准确率为90.91％,能够对学生的编程能力进行评估.
Abstract
Aiming at the problem that programming ability is difficult to evaluate in traditional teaching,a method of evaluating students'programming ability using BP neural network is proposed.The weight of each index of programming ability is obtained by combining entropy weight method and deviation maximization meth-od,and the comprehensive score of students'programming ability is calculated to optimize the input of the eval-uation model,and the input space of the model is increased by adjusting the number of hidden layers.In order to reduce the output error of the evaluation model and compare the evaluation effect of different BP neural net-works,the experimental results show that when the number of hidden layers is 2,the output error of the evalu-ation model is the smallest,and the prediction accuracy of the model is 90.91％,which can evaluate the students'programming ability.
关键词
神经网络算法/BP神经网络/教育数据挖掘/编程能力Key words
neural network algorithm/BP neural network/educational data mining/programming ability引用本文复制引用
基金项目
湖北省湖北师范大学重点项目(2022007)
出版年
2024
NETL
NSTL
万方数据