基于模糊评价与BP神经网络模型的耕地质量评价对比研究——以淮北平原凤台县为例

Comparison of Fuzzy Evaluation and BP Neural Network applied for Cultivated Land Quality Evaluation——A Case Study at Fengtai County

以凤台县为例,提取耕地,在Arc GIS中通过基础矢量图件的叠置分析建立耕地质量评价单元。选取与耕地质量息息相关的评价指标,分别建立模糊模型和BP神经网络模型。再用建好的模型对耕地质量评单元进行评价,将其评价结果从数量和空间上与依据真实常年单产划分的耕地质量等级进行对比,发现在数量上BP神经网络模型评价结果的相对误差均比模糊模型评价结果的相关误差小,在空间分布上BP神经网络模型评价结果也比模糊模型评价结果更接近于真实分布。最后绘制空间分布图,直观展示不同评价方法分布结果。

The Fuzzy model and BP neural network model were established firstly, and then the evaluation results from the two models were compared to the real perennial division yields from the quantity and space. The relative error of BP neural network model are smaller than fuzzy model. The spatial distribution of evaluation results of BP neural network model are closer to the real distribution of evaluation results. Finally spatial distribution maps were drawn by using visual display of the distribution of the results of different evaluation methods.