基于GA-BP神经网络算法的马铃薯晚疫病预测模型

摘要

Based on genetic algorithm of BP neural network,we built the potato late blight forecasting model. After the normalization processing of original samples,genetic algorithm was applied to optimize the BP neural network structure,initial weights and thresholds,and then BP neural network training was used to build the potato late blight prediction model. By using the genetic algorithm,the model improves the defects of BP neural network algorithm itself,improves the learning accuracy and the prediction accuracy. The simulation results show that,GA-BP neural network model prediction accuracy is higher,with lower error rate and better stability. The GA-BP neural network algorithm applied to potato late blight prediction model is feasible,can achieve the fast prediction of popularity.%应用基于遗传算法的BP神经网络构建马铃薯晚疫病预测模型，对原始样本进行归一化处理，应用遗传算法优化BP神经网络的结构、初始权值、阀值，通过BP神经网络训练构建马铃薯晚疫病预测模型，利用遗传算法来改善BP神经网络算法本身的缺陷，提高学习精度，预测准确度。仿真结果表明，GA-BP神经网络模型预测准确度较高，误差率较低，稳定性较好。实践证明，将GA-BP神经网络算法应用于马铃薯晚疫病预测模型中是可行的，能够实现晚疫病流行程度的快速预测。