Exploratory geospatial data analysis using self-organizing maps

Abstract

The rapidly increasing volume of digital geographic data is overwhelming for conventional analysis techniques and methods. Therefore new approaches are needed to transform data into information, and ultimately, into knowledge. Exploratory data analysis is a foundation stone in this process. It is concerned with the formation of a simplified overview of data sets. Clustering and projection are among the examples of useful methods to achieve this task. The Self-Organizing Map (SOM) algorithm performs both, in a non-linear mapping from a high-dimensional data space to a low-dimensional space aiming to preserve the topological relations in the data. The aim of this thesis is to demonstrate the effectiveness of SOM application in visual exploration of physical geography data to support the delineation of Portuguese mainland regions. The main justifications for the application of SOM in this issue are its features of stressing local factors and topological ordering. For experimental assessment, the public domain thematic maps from Instituto do Ambiente are used. Several authors’ maps of Portuguese regions are used for evaluation of empirical results.(...)