基于BP神经网络算法的成绩预测模型研究

基于sigmoid激活函数，建立了一种BP神经网络模型。通过对某高中2006年至2015年间的高考平均数据样本进行学习，修正了权值和阈值。系统最大相对误差为0.22%，关联度为0.6667，小误差概率为0.98，方差比为0.0002，预测结果精度为高。用于2016年至2020年间该校高考平均成绩的预测中发现，预测结果与实际结果的最大绝对误差仅为2分。对该校2021年的高考平均成绩进行了预测，最终预测结果为571分。

Research on Performance Prediction Model Based on BP Neural Network Algorithm

Based on sigmoid activation function, a BP neural network system was established. Its weights and thresholds were further revised by learning the average data of college entrance examination from 2006 to 2015 of a certain high school. The maximum relative error of the system was 0.22%, the correlation degree was 0.6667, the probability of small error was 0.98, and the variance ratio was 0.0002. The prediction accuracy of this BP neural network system was high. In the prediction of the average score of the college entrance examination from 2016 to 2020, it was found that the absolute error between the predicted result and the actual result was only 2 points. In addition, the BP neural network also was used to predict the average score of the college entrance examination in 2021, and affording a final prediction result as 571 points.