一种分组模式下的土壤重金属含量预测模型

A Prediction Model of Soil Heavy Metal Content Based on Grouping Mode

针对传统预测模型在土壤重金属含量预测上表现不佳问题,以土壤采样点数据集中的经度、纬度、高度以及农作物类型作为输入变量,建立一种基于分组教学优化算法的分组模式预测模型(GTOA-BP)。对武汉新城区土壤采样数据进行仿真预测,将GTOA-BP模型与BP神经网络和径向基神经网络模型进行实验比较,结果表明:GTOA-BP的4种误差数据均低于其他两种模型。与BP神经网络相比,GTOA-BP的MAPE和SMAPE分别下降了9.97%和8.86%,与径向基神经网络相比,GTOA-BP的MAPE和SMAPE分别下降了6.24%和5.97%,说明该模型能降低神经网络训练的误差,提高预测精度。

In view of the poor performance of traditional prediction model in soil heavy metal content prediction and a few soil heavy metal content prediction models,a prediction model of grouping mode based on grouping teaching optimization algorithm is established by taking longitude,latitude,height and crop type as input variables.Based on the soil sampling data of Wuhan new urban area,the simulation prediction is carried out,and the experimental results are compared with BP neural network and radial basis function neural network.The experimental results show that the four error data of GTOA-BP are lower than those of the other two models.Compared with BP neural network,the MAPE and smape of GTOA-BP are decreased by 9.97%and 8.86%respectively,and compared with radial basis function neural network,the MAPE and SMAPE of GTOA-BP are decreased by 6.24%and 5.97%respectively.The results illustrate that the model can reduce the error of neural network training and improve the prediction accuracy.