基于模糊预测对象的证据权方法及其在土地沙漠化评价中的应用

Weights of Evidence Method Based on Fuzzy Training Layer and Its Application in Desertification Assessment

为了使在地理信息系统中被广泛用于点事件预测的证据权方法能对面事件进行评价和预测,提出了一种新的基于模糊训练层的证据权方法.它是一种更广泛的证据权方法,与普通证据权方法所不同的是,它的训练层是模糊集合,其取值是它的隶属度.通过适当的变换也可以把点训练层转换为模糊集合.因此,该方法可以对面事件、点事件和线事件进行评价和预测.该方法可以处理训练层和证据层均为模糊集合的情况,被称为双重模糊证据权方法.作为该方法的一个应用实例,本文介绍毛乌素沙漠边缘的晋陕蒙地区土地沙漠化评价的应用实例.

This paper proposes an extended version of the fuzzy weights of evidence method, based on the fuzzy training layer for assessing and predicting the occurrence of area-events. It can be considered as a generalization of the ordinary weights of evidence method, which is used to predict the occurrence of point-events with known evidences and training layers. Unlike the ordinary weights of evidence using point training sets, the new method involves training data as a fuzzy set. Point training data can be converted to a fuzzy set, therefore the new method can predict both point-events and area-events. When both evidence and training data are fuzzy sets, the new method acts as a dual fuzzy weights of evidence method. A case study has been used to demonstrate the application of the method in assessing desertification in the area in the contact of Shanxi and Shaanxi Provinces and Inner Mongolia Autonomous Region.