作　　者： ; ; ; ;

机构地区： 华南农业大学信息学院

出　　处： 《华南农业大学学报》 2005年第3期108-110,共3页

摘　　要： 将广东省土壤资源类型图和各种评价因子的单要素图层进行叠置分析,以土壤资源类型图的图斑作为土壤质量评价单元,从各生成图层的PAT文件中提取高程、坡度、土壤有机质含量、土壤质地、土壤pH、土壤利用类型、地貌类型和土壤类型等评价属性,用决策树方法预测土壤质量等级,并以定量规则方式表达所获取的知识.结果表明,其知识表达易于理解,预测准确率为96.61%. This paper overlayed the coverage of land-resource-type map with single-factor-coverages of each evaluation factor separately. The spots on land-resource-type map were studied as evaluation units for soil quality,attributes were extracted from the polygon attribute table (PAT) of each resulted coverage, such as elevation, slope, soil organic material content, soil texture,soil pH, land uses, landform type and soil type. Decision tree is used to predict soil quality grade, the knowledge are expressed in quantitative rules. The results showed that it was easy to understand the knowledge,and accuracy could be 96.61%.

关 键 词： 空间数据挖掘 决策树 土壤质量 地理信息系统

领　　域： [自动化与计算机技术] [自动化与计算机技术]