|
|
|
Abstract
Geographic Data Analysis and System Modelling Toolbox using Soft-Computing and Non-Gaussian Methods
R. Sivakumar
Reader
Birla Institute of Technology, Mesra, Ranchi,
India
Email: skm_ram@yahoo.com
Pradeep Mohan
Department of Computer Science and Engineering
Birla Institute of Technology
Mesra
Ranchi
Email: pradeep.mohan.cs@gmail.com
Anish Mitra
Department of Electrical and Electronics Engineering
Birla Institute of Technology
Mesra
Ranchi
Email: mitra.anish@gmail.com
The inherent complexity and large size of geographic data sets necessitates the use of advanced analysis and optimization techniques for performing various analytical tasks on geographic data sets. The various tasks that could be performed in the domain of Geographic Data Analysis are statistical analysis, data clustering, spatial data indexing, visualization, data analysis, system modelling, function estimation, parameter evaluation, dimensionality reduction, Soft Computing based System modeling procedures and Independent Component Analysis (ICA) techniques. The absence of any toolbox or tools provides the motivation for the present task. This paper analyses the need for such a tool box designed in a tool like Matlab. Further, the paper presents the layout and design procedure adopted to design such a toolbox. The paper also introduces various advanced data analysis techniques which have been integrated as functions into the proposed toolbox. Also, the importance of superior, robust and high performance techniques using evolutionary algorithms is reported. The paper also provides a case study and demonstrates the various functions and their performance for a particular Geo-Data analysis and system modeling problem of an arbitrary study area.
Keywords: Geographic Data Analysis, Parameter evaluation, clustering, hybrid dimensionality reduction, Evolutionary Optimization, System Modeling
|
|
|