Comparison of Predictive Data Mining Methods for their Application in a Design Process
Year: 2010
Editor: Andreas Dagman; Rikard Söderberg
Author: Röhner, Sebastian; Donhauser, Maximilian; Wartzack, Sandro
Section: Virtual Product Realization
Page(s): 365-374
Abstract
A design process is characterized by acquisition and processing of information or knowledge, respectively. The acquisition of information and knowledge can be direct, indirect or automatic. This paper presents an approach of how to automatically extract design knowledge from process data. Several data mining algorithms were tested on a benchmark data sets to find the best performing ones. Especially the C&RT algorithm turned out to be interesting as its extracted knowledge is rule-represented and can be used to cope with tasks that up to now require the experience of expert designers.
Keywords: artificial intelligence, data mining