论文部分内容阅读
将支持向量机方法(Support Vector Machines,SVM)应用于北京平原地区粮田土壤养分肥力评价,与判别分析评价的结果相比较,支持向量机能够较好的揭示研究区域土壤养分肥力的现状。以专家评判为准,判别准确率可达84.62%;其中中等养分肥力以上的样本占58.19%,表明北京平原地区粮田土壤养分肥力的总体水平处于中等偏上。根据支持向量机评价的结果进行Kriging最优内插,并绘制出土壤养分肥力等级图,对各区县肥力情况进行分析,表明房山、海淀、丰台和门头沟属较高肥力区县,怀柔和平谷处于中等水平,其他区县养分肥力水平则较低。
Compared with the results of discriminant analysis and appraisal, Support Vector Machines (SVM) can be applied to evaluate the soil nutrient fertility of grain fields in Beijing Plain. Support vector machines (SVMs) can better reveal the status of soil nutrient fertility in the study area. Judging by experts, the accurate rate of identification can reach 84.62%. Among them, 58.19% of the samples are above medium nutrient fertility, indicating that the overall level of soil nutrient and fertility of grain fields in the plain of Beijing is above average. According to the results of support vector machine (SVM), Kriging optimal interpolation and soil fertility rank map were drawn, and the fertility of each county was analyzed, which showed that Fangshan, Haidian, Fengtai and Mentougou were in high fertility counties and Huairou and Pinggu were in Medium level, other districts and counties have lower levels of nutrient fertility.