Design and Development of Software Framework for Spatio-Temporal Data Analysis and Mining

Abstract

Space and time are the basic dimensions of our existence and thus needed to be explored in order to reveal hidden information and useful knowledge. Spatiotemporal data are generated in various fields. Such as agriculture, defence, meteorology, crop sciences, medicine, transportation etc. These datasets are growing rapidly with the use of geographic location based devices and satellite images. Data are captured in various formats at multiple levels of granularity both in space and time. The increased volume of datasets demand research in the field of spatiotemporal data mining and analysis.

In the thesis, a framework has been proposed and developed for the analysis and mining of spatio-temporal data. It has been designed as two subsystems i.e. Automated and Customized. Automated system is meant for multi-dimensional analysis and visualization. A Data warehouse is also created for the use of Automated System. For performing all the data mining tasks manually, customized system has been developed, which is an extension of open source data mining tool, WEKA.