Automated Visual Clustering Functionality for Improved and Effective Mining of GIS Data

Abstract

The technical progress in computerized data acquisition and storage has resulted in the growth of vast databases for Geographical Information Systems. This has led to continuous increase and accumulation of huge amounts of the computerized data that have far exceeded human ability to completely interpret, analyze and use. In order to understand and make full use of these data repositories, various techniques have been put forward. However, these techniques are not fully reliable as they are not as efficient or of high performance as is expected. This thesis attempts to improve on the efficiency of existing spatial data mining techniques to ensure more efficient and high performance spatial data mining functionality in the present framework and tools used for spatial data mining. This will be done by integrating various techniques with available technologies. The focus of this project is on improving performance and efficiency of spatial clustering, one of the commonly used spatial data mining methods by integrating visualization into clustering with an aim to provide an interactive, efficient and user-friendly approach to this important process for GIS data.