基于因素法和SVM模型的耕地质量评价方法研究

Cultivated Land Quality Assessment Methodology Based on Factor Method and SVM Model

为减少耕地质量评价工作中确定评价指标权重人为因素的影响,选取江西省奉新县为研究对象,引入机器学习技术,建立基于SVM模型的耕地质量评价模型,探索客观、高效的耕地质量评价方法。本文应用因素法和SVM模型法分别对奉新县耕地质量等别进行划定,通过对比分析两种方法的评价结果,对两种方法的优劣及适用条件进行了判定。结果表明,与因素法相比,SVM模型法具有等别划分高效的优势;SVM模型法耕地质量评价结果的正确率高达96.93%,符合耕地质量评价精度要求。应用SVM模型进行耕地质量评价具有测算高效、受人为因素干扰小等优点,可以广泛应用于耕地质量等别更新、新增耕地质量等别划定等工作。

In order to reduce the human factor influence on the weight of evaluation factors by the traditional evaluation method and explore an objective and efficient method of cultivated land quality,the evaluation model of cultivated land quality in Fengxin,Jiangxi Province was established based on SVM model and by introducing machine learning.The factor method and SVM model were used to classify cultivated land quality in Fengxin.Compared with the factor method,determined the advantages and disadvantages of the two methods the applicable conditions.The results show that the SVM model method has the advantage of high efficiency in classification and the accuracy of cultivated land quality evaluation by using SVM model was higher than 96.93%,which was in line with the accuracy requirement of cultivated land quality evaluation.The SVM model the advantages of high efficiency and little interference by human factors and can be widely applied in the update of cultivated land quality level and the classification for new cultivated land quality and so on.