Combination of designed immune based classifiers for ERP assessment in a P300-based GKT
Moradi, Mohammad Hassan
Date of Issue2012
School of Electrical and Electronic Engineering
Constructing a precise classifier is an important issue in pattern recognition task. Combination the decision of several competing classifiers to achieve improved classification accuracy has become interested in many research areas. In this study, Artificial Immune system (AIS) as an effective artificial intelligence technique was used for designing of several efficient classifiers. Combination of multiple immune based classifiers was tested on ERP assessment in a P300-based GKT (Guilty Knowledge Test). Experiment results showed that the proposed classifier named Compact Artificial Immune System (CAIS) was a successful classification method and could be competitive to other classifiers such as K-nearest neighbourhood (KNN), Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM). Also, in the experiments, it was observed that using the decision fusion techniques for multiple classifier combination lead to better recognition results. The best rate of recognition by CAIS was 80.90% that has been improved in compare to other applied classification methods in our study.
Research journal of applied sciences, engineering and technology
© 2012 Maxwell Scientific Organization. This paper was published in Research Journal of Applied Sciences, Engineering and Technology and is made available as an electronic reprint (preprint) with permission of Maxwell Scientific Organization. The paper can be found at the following official URL: [http://www.maxwellsci.com/jp/abstract.php?jid=RJASET&no=211&abs=23]. One print or electronic copy may be made for personal use only. Systematic or multiple reproduction, distribution to multiple locations via electronic or other means, duplication of any material in this paper for a fee or for commercial purposes, or modification of the content of the paper is prohibited and is subject to penalties under law.