您的位置:首页  > 论文页面

基于贝叶斯最大熵与历史数据的土壤属性空间预测

发表时间:2012-10-15  浏览量:1668  下载量:853
全部作者: 杨勇,李卫东,贺立源
作者单位: 华中农业大学资源与环境学院
摘 要: 以武汉市汉南区2005年和2009年土壤有机质分布和样点数据为研究对象,基于累计概率分布函数从2005年土壤有机质分布状况中获取软数据,将2009年的土壤采样数据作为硬数据,利用贝叶斯最大熵(Bayesian maximum entropy, BME)方法将两类数据综合起来用于土壤属性空间预测,通过与不同样点密度下的经典地统计方法对比,表明所提方法有较高的预测精度,特别是在样点数据较少的情况下,BME方法的优势更为明显。最后对BME方法的优势和不足进行评价。�
关 键 词: 土壤学;贝叶斯最大熵;软数据;土壤属性;空间预测�
Title: Spatial predicition of soil properties based on Bayesian maximum entropy and historical data
Author: YANG Yong, LI Weidong, HE Liyuan
Organization: College of Resources and Environment, Huazhong Agricultural University
Abstract: In this paper, soil organic matter distribution of 2005 and soil sampling of 2009 in Hannan, Wuhan were used as experimental data, and Bayesian maximum entropy (BME) was used to estimate spatial distribution of soil continuous properties. The soft data was calculated with soil organic matter distribution of 2005 based on cumulative probability distribution function, and soil was sampling data of 2009 was used as hard data. The prediction performance of BME was compared with that of Kriging under different densities. It was found that BME was more accurate, especially in the case of less sample points, the advantages of the BME method were more obvious. Finally, The advantages and disadvantages of the BME method was discussed.�
Key words: soil science; Bayesian maximum entropy; soft data; soil properties; spatial prediction
发表期数: 2012年10月第19期
引用格式: 杨勇,李卫东,贺立源. 基于贝叶斯最大熵与历史数据的土壤属性空间预测[J]. 中国科技论文在线精品论文,2012,5(19):1864-1870.
 
0 评论数 0
暂无评论
友情链接