土壤图更新中基于土壤类型面积分级的训练样点选择方法

发布时间：2018-03-25 03:08

本文选题：训练样点　切入点：数据挖掘模型　出处：《土壤学报》2017年01期

【摘要】：基于数据挖掘模型的土壤图更新是一项重要的研究。数据挖掘模型构建中训练样点的质量不仅决定其对研究区土壤-环境关系表达的充分程度,而且会对推理制图的结果产生至关重要的影响。本文提出一种基于土壤类型面积分级的典型训练样点选择方法,即依据土壤面积对土壤类型分级,并按照等级之间的比例关系基于典型点选择训练样点。将方法应用于更新美国威斯康星州Raffelson流域的传统土壤图,并与另外两种训练样点选择方法对比,以验证该方法的有效性。结果表明,500次重复实验中,本研究方法与另外两种训练样点选择方法相比,能够更新传统土壤图的比例分别为79.5%、71.8%和63.6%,而且其推理制图结果更符合研究区土壤分布的特征。本研究所提方法是一种有效的训练样点选择方法。
[Abstract]:The updating of soil map based on data mining model is an important research. The quality of training sample points in data mining model construction not only determines the sufficient degree of expression of soil-environment relationship in the study area. In this paper, a typical training point selection method based on soil type area classification is proposed, which is based on soil area classification. The method is applied to update the traditional soil map of Raffelson watershed in Wisconsin, and compared with the other two training sample selection methods. In order to verify the effectiveness of this method, the results show that the proposed method is compared with the other two training sample selection methods in 500 repeated experiments. The proportions of the traditional soil maps were 71.8% and 63.6%, respectively, and the inferential mapping results were more consistent with the characteristics of soil distribution in the study area. The method proposed in this study is an effective training sample selection method.
【作者单位】：南京师范大学地理科学学院;资源与环境信息系统国家重点实验室(中国科学院地理科学与资源研究所);虚拟地理环境教育部重点实验室(南京师范大学)江苏省地理环境演化国家重点实验室培育建设点江苏省地理信息资源开发与利用协同创新中心;Department
【基金】：国家自然科学基金项目(41431177;41471178) 江苏省高校自然科学研究重大项目(14KJA170001) 江苏省高校研究生科研创新计划项目(KYLX15_0715) 国家重点基础研究发展计划973项目(2015CB954102) 千人计划资助~~
【分类号】：S159

【相似文献】