Affordable Access

Tree Induction of Spatial Choice Behavior

  • Computer Science
  • Design
  • Mathematics


A282 Tree Induction of Spatial Choice Behavior Jean-Claude Thill Department of Geography State University of New York at Buffalo Amherst, New York 14261, USA Aaron Wheeler Department of Computer Science University of New Mexico Albuquerque, New Mexico 87131, USA Acknowledgement. The first author wishes to acknowledge partial support of the National Science Foundation to the National Center for Geographic Information and Analysis at SUNY/Buffalo under award SBR-9600465. 2 1. Introduction Machine learning, a branch of artificial intelligence, investigates the mechanisms by which knowledge is acquired through experience. A large number of machine learning methods and algorithms have been developed, including neural computing (Freeman and Skapura, 1991), case- based reasoning (Kolodner, 1993), genetic algorithms (Goldberg, 1989), and inductive learning (Quinlan, 1988). These approaches form the essential toolbox of methods to extract useful information from data sets built into the knowledge base of expert systems. It has been argued that these computational methods are not only useful for the design and implementation of effective and efficient decision support and expert systems, but also as support tools in furthering scientific knowledge discovery above and beyond what conventional methods of inquiry have so far permitted. In the domain of the Spatial Sciences, this viewpoint is forcefully advocated in the research white paper on "Spatial Analysis in a GIS Environment" of the University Consortium for Geographic Information Science (UCGIS, 1997). In this chapter, we discuss the merit of inductive learning as an analysis tool in spatial decision making theory. We analyze the capability and applicability of Ross Quinlan's (1993) C4.5 decision tree induction algorithm to the class of problems involving the choice among travel destination within an urban area. The chapter reviews the relevant destination choice modeling literature, des

There are no comments yet on this publication. Be the first to share your thoughts.


Seen <100 times