基于模糊关系识别的多要素空间离散化方法——以江苏阜宁人口与经济分析为例

MULTI-ELEMENT SPATIAL DISCRETIZATION METHOD BASED ON FUZZY RELATIONS RECOGNITION——A Case Study of Funing County,Jiangsu Province

社会、经济要素的空间离散化是精细化县级主体功能区划的重要需求,本文提出了一种基于模糊关系识别的空间数据离散化方法。该方法利用广义权距离实现对专家知识与多系统分层要素的综合集成,并通过建立待离散化要素与其影响指标的模糊关系识别模型,获得空间离散化权重。以2009年江苏省阜宁县人口、GDP以及经济发展水平的空间离散化为例进行实例分析。结果显示,本文提出的空间离散化方法具有较好的准确性与可信度,可较好揭示各影响要素对待离散要素的空间影响。

Traditional social-economic data acquisitions are primarily based on the administrative divisions,and taking into account the expert knowledge to build a new multi-element spatial data affect the discrete method is an important way to enhance the county-level divisions of the main function planning.However,lots of modern main functional zoning planning requires high resolution and precision spatial data for model analysis and decision support.Hence,spatial discretization of social and economic indexes is an important requirement for the main functional zoning planning.Lots of existing methods such as linear regression,kernel smoothing etc.have already been developed.Existing spatial discretization methods are usually affected by the diversity,complexity,and spatial heterogeneity of observed data,which cannot achieve well balance among multiple factor integration,the accuracy and precision and the integration of expert empirical knowledge.To fulfill the gap between the spatial discretization method and the main functional zoning planning needs.This paper presents a multi-element spatial discretization method based on fuzzy relations recognition.Five subsystems,such as land usage,terrain conditions,road traffics,ecological,spatial factors etc.are considered to construct the affection factor index system.The weight of the affect factors are determined by the expert choice with AHP.By defining and applying a generalized weight distance,the expert＇s knowledge is integrated.Through the establishment of discrete elements to be indicators of the fuzzy relation and by building its influence identification model,we can access to space discretization weights.The results show that the proposed space discretization method has better accuracy and reliability,and can better reveal the spatial impact of the discrete elements and their affect factors.