Authors: D.V. Chandra Shekar, V. Sesha Srinivas, J. Pratap Reddy
Addresses: Department of Computer Science, TJPS College (PG courses), Guntur, Andhra Pradesh, India. ' Department of Computer Applications, RVR and JC Engineering College, Guntur, Andhra Pradesh, India. ' Department of Computer Applications, St. Anns College, Guntur, Andhra Pradesh, India
Abstract: The advances in manufacturing technology and proliferation of information from research laboratories have posed the problem of information explosion. To understand any system|s behaviour it would be significant if the process can be visualised as a mapping from inputs to outputs. In materials research also, the generation of hypothesis about various materials properties and the compositions leading to such properties can be better understood by multivariate visualisation and machine learning techniques. Attempts have been made to develop a novel method for multivariate visualisation using parallel coordinate systems. Using two popular methods, clustering and the k-nearest neighbour technique, solves the materials classification problem. The prototype will allow the user to interactively discover the hidden patterns by semi-automatic generation of hypothesis and testing. As a case study the developed prototype has been tested on an aluminium alloy database.
Keywords: knowledge discovery; visual data mining; multivariate visualisation; parallel coordinates; manufacturing; materials research; clustering; k-nearest neighbour; materials classification; aluminium alloys.
International Journal of Mechatronics and Manufacturing Systems, 2010 Vol.3 No.1/2, pp.131 - 143
Available online: 02 Dec 2009 *Full-text access for editors Access for subscribers Purchase this article Comment on this article