Modeling Urban Growth with Geographically Weighted Multinomial Logistic Regression

摘要：

Spatial heterogeneity is usually ignored in previous land use change studies.This paper presents a geographically weighted multinomial logistic regression model for investigating multiple land use conversion in the urban growth process.The proposed model makes estimation at each sample location and generates local coefficients of driving factors for land use conversion.A Gaussian function is used for determine the geographic weights guarantying that all other samples are involved in the calibration of the model for one location.A case study on Springfield metropolitan area is conducted.A set of independent variables are selected as driving factors.A traditional multinomial logistic regression model is set up and compared with the proposed model.Spatial variations of coefficients of independent variables are revealed by investigating the estimations at sample locations.

关键词： geographically weighted multinomial logistic regression land use conversion urban growth GIS

作者: Jun Luo Nagaraj Kapi Kanala

作者单位: Missouri State University,901 S National Avenue,Springfield,MO,USA 65897

会议类型: 国际会议

会议名称: 第16届国际地理信息科学与技术大会(16th International Conference on GeoInformatics and the Joint Conference)

会议地点: 广州

会议语种:英文

在线出版日期: 2008-06-28（万方平台首次上网日期，不代表论文的发表时间）