摘要
在探讨应用空间数据挖掘的决策树方法于土地适宜性评价的研究思路后,以琼海市的3个乡镇作为试验区,选取了8个指标(影响因子)作为测试属性,共提取了67个样本,运用C4.5算法构建土地适宜性评价决策树并进行了分析,表明所得规则符合实际且易于理解。然后运用规则对试验区土地进行评价,并与应用其它评价方法所得结果作对比分析。结果表明:基于决策树的土地适宜性评价方法不依赖于经验知识,得到的评价结果具有较高准确率,能够满足评价的要求。
The discussion of research process of land use suitability based on decision tree model is followed by a selection of three towns in Qionghai City as a test case, in which 8 factors are chosen as testing attributes, and 67 land parcels are collected as samples. Then the authors build a decision tree for suitability evaluation of land use by C4.5 program with these data, and get the rules from it, discovering that it accords with the reality and can be understood easily. In the end, the authors use the rules to evaluate the land in the tested region and compare it with some other results computed with other methods, with the result that the decision tree model is not dependent on experience, and therefore can be applied in suitability evaluation of land use with good result.
出处
《国土资源科技管理》
2007年第3期21-25,共5页
Scientific and Technological Management of Land and Resources
关键词
决策树
土地适宜性评价
智能评价
Decision Tree model
suitability evaluation of land use
intelligent evaluation