RBF神经网络的土壤养分肥力评价研究

Research on Soil Nutrient Fertility Evaluation of RBF Neural Network

土壤养分肥力等级是土壤特征的综合反映。以黑龙江850农场为研究区域,应用RBF神经网络方法构建该区的土壤养分肥力评价模型。以土壤养分指标作为神经网络的输入,土壤养分等级作为输出。通过实验模拟训练和预测,RBF神经网络取得的结果较好,与相关的BP神经网络模型进行实验比较,实验结果表明RBF神经网络得到的等级结果精度更高,该模型的建立与预测为土壤养分肥力等级评价提供了新途径。

Levels of soil fertility was comprehensive reflection of soil characteristics. Using 850 farm of Heilongjiang as study area,the RBF neural network method was used to construct soil fertility evaluation model of this area. This model used soil nutrient index as input of neural network and soil nutrient levels as the output. Through the experimental simulation training and prediction,RBF neural network had better results. Compared with BP neural network model,the experimental results showed that RBF neural network had the higher accuracy of ranking results,So it provided the new path of evaluation and prediction model for soil fertility levels.