• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International symposium on spatial accuracy assessment in natural resources and environmental science >Cartographic Delineations Using Variables Selected by Principal Component Analysis (PCA) and Semivariograms, and Map Validation by Continuous Classification
【24h】

Cartographic Delineations Using Variables Selected by Principal Component Analysis (PCA) and Semivariograms, and Map Validation by Continuous Classification

机译：使用主成分分析（PCA）和半啮盘函数选择的变量的制图描绘，并通过连续分类来映射验证

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Many variables are obtained in the standard sampling process to get information to delineate map units. Some of these could not add useful information because having minimal variability; the line of a unit based on these properties could be traced at any direction, or variables with a high variability at excessive short range force to delineate map units with a surface so minimal that could be not manageable. In this work, a set of 115 soil observations with measures of 14 properties, between 0 to 20 cm and 40-60 cm deep, were reduced in dimensionality discarding by PCA the properties with little variation, and discarding the variables with short-range variation analysing the scatter diagrams (SD) and the range of the semivariogram (SVA). The reduced set not discarded by PCA, SD and SVA consist of sand and clay at both deep and electrical conductivity between 40 to 60 cm deep. Using multivariate analysis with the original and interpolated by kriging data, the map units delineated using the reduced set performs better than use all the variables, or the variables discarded by the other methods. With independent observations, was proved the predictive value of the proposed units by continuous classification.
机译：在标准采样过程中获得了许多变量，以获取信息以解释地图单元。其中一些不能添加有用的信息，因为具有最小的变异性;基于这些属性的单元的线可以在任何方向上追踪，或者在过度短的范围力下具有高可变性的变量，以用表面划分地图单元，使得可能无法管理的表面。在这项工作中，通过多维变化的性能丢弃的维度丢弃，一系列具有14个性能的115个土壤观察，其尺寸为0至20厘米和40-60厘米的尺寸。分析散点图（SD）和半啮盘仪（SVA）的范围。 PCA，SD和SVA没有丢弃的减少的组成的砂和粘土在深度和电导率下，在40至60厘米的深处。使用具有kriging数据的原始和插值的多变量分析，使用缩小组描绘的地图单元比使用所有变量更好地执行，或者其他方法丢弃的变量。通过独立观察，通过连续分类证明了提出单位的预测价值。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号