Home // COGNITIVE 2012, The Fourth International Conference on Advanced Cognitive Technologies and Applications // View article

Speed Up Learning in a Test Feature Classifier Using Overlap Index Lists

Authors:
Yoshikazu Matsuo
Takamichi Kobayashi
Hidenori Takauji
Shuni'ichi Kaneko

Keywords: Test Feature Classifier; Overlap Index List; Speed Up Learning; Curse of Dimensionality

Abstract:
This paper presents a novel low cost learning algorithm for a Test Feature Classifier using Overlap Index Lists (OILs). In general, pattern classifiers require a large amount of training data to attain high performance, which is expensive in terms of computation time. Our proposed algorithm uses OILs to efficiently find and check combinations of features starting with lower dimensions and working up-to higher ones. Our algorithm can solve classification problems in real industrial inspection lines with large reductions in computation time.

Pages: 63 to 66

Copyright: Copyright (c) IARIA, 2012

Publication date: July 22, 2012

Published in: conference

ISSN: 2308-4197

ISBN: 978-1-61208-218-9

Location: Nice, France

Dates: from July 22, 2012 to July 27, 2012