Estimating spatial logistic model: A deterministic approach or a heuristic approach?

• Published in: Information sciences Vol. 330; pp. 358 - 369
• Main Authors: Zhang, Xinxin, Huang, Bo, Tay, Richard
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 10.02.2016
• Subjects: Estimating; genetic algorithm; Genetic algorithms; Geographic information systems; Heuristic methods; land use; Logistic regression model; Logistics; Mathematical models; Newton-Raphson method; Regression; Satellite navigation systems; spatial sampling and filtering; genetic algorithm; spatial sampling and filtering; Newton-Raphson method; land use; Logistic regression model
• ISSN: 0020-0255; 1872-6291
• DOI: 10.1016/j.ins.2015.10.015

This paper evaluates the performance of a deterministic method (Newton–Raphson, NR) and a heuristic method (Genetic Algorithm, GA) for solving the maximum likelihood estimation in spatial logistic analysis. A spatial logistic regression model equipped with a filtering process is formulated to examine the relationship between land use change and various spatial determinants such as population density, distance to road, distance to commercial center and neighborhood characteristics. Geographic Information System (GIS) is used to acquire spatial samples and perform spatial analysis. The NR method and GA are utilized, respectively, to estimate the coefficients that maximize the likelihood of the spatial logistic regression model. Both methods are compared in terms of the maximum likelihood and computing time. The experimental results show that the NR method can achieve a better likelihood and is also much faster than the GA method. Therefore, the NR method is recommended for estimating a spatial logistic regression model although GA can also be employed.