ABSTRACT
This chapter discusses the conceptualisation of predictive modelling in spatial data analysis (SDA). A decision process for spatial data can have one or more of the following purposes: assessment, search for patterns, understanding of natural or social phenomena, management of information, and prediction of location and intensity of the phenomena under study. Techniques for SDA were developed in the 1960s and 1970s when the main concern in geographic information systems applications was in developing infrastructures and facilities to store and manage data on resources, and less in the analysis of the data. The greatest difficulty in SDA is represented by the complexity of formulating a conceptual model of a process into a computational form. A discussion of such difficulties by A. A. Green and M. Craig highlights the need to develop a rationale to associate different types and layers of information within a multiple dataset.
