基于模拟退火优化BP神经网络的pH值预测

Optimized BP Neural Network Based on Simulated Annealing Algorithm for pH Value Prediction

为自动准确测定水质p H值,采用大量的具有代表性的p H值检测数据为样本,提出了一种基于模拟退火优化BP神经网络的p H值预测方法。利用模拟退火算法优化BP网络的权值,调整优化样本的选取和隐层神经元数,训练BP神经网络预测模型得到最优解。由测试样本对网络进行了预测试验,并与非线性回归的预测结果进行了对比。结果表明,该方法对水质p H值预测具有较好的非线性拟合能力和更高的预测准确性。

In order to determine accurate pH value of water automatically,sufficient and typical data of pH value measuring tests are collected as samples,and a pH value prediction method of optimized BP neural network based on simulated annealing algorithm is presented. The simulated annealing algorithm is employed to optimize the weights and thresholds of BP neural network,and selection methods of the training samples and the number of the hidden layer nodes are improved,thus yield an optimal solution of BP neural network. The obtained BP neural net-work is tested by samples,and the prediction results are compared with ones given by a nonlinear regression method. Experimental results exhibit that the proposed method provides better fitting ability and higher accuracy for pH value prediction.