基于遗传算法优化BP网络的高校生师比预测模型分析

Analysis of Forecast Model for University Students/Teacher Ratio Base on Genetic Algorithm Optimization of BP

标准BP算法采用的是非线性无约束极值问题求解方法中最古老又十分基本的方法 -梯度法(梯度下降法)。标准BP算法具有学习效率低,收敛速度慢,容易陷入局部极小点。通过标准BP算法模型和遗传算法优化的BP算法模型对高校生师比的预测结果进行比较。结果表明,遗传算法优化的BP神经网络的权值和阈值具有良好的泛化能力,提高了高校生师比预测精度和效率。

The standard of the BP algorithm is nonlinear unconstrained extreme problem solving method of the most ancient and very basic methods-gradient method （ gradient descent method）. The BP algorithm has the character of low learning efficiency, slow convergence speed, which could easily get into the local minimum points. The standard of the BP algorithm model and genetic algorithm optimize the BP algorithm model of university teacher/ student ratio comparing the prediction result is set up. Simulations show that the genetic algorithm optimizes the BP neural network of weights and threshold value has good generalization ability, which improve the university student/teacher ratio in prediction accuracy and efficiency.