A COMPARISON BETWEEN THE PROPORTIONAL KEEN APPROXIMATOR AND THE NEURAL NETWORKS LEARNING METHODS

Peyman Kabiri

Abstract

The Proportional Keen Approximation method is a young learning method using the linear approximation to learn hypothesis. In the paper this methodology will be compared with another well-established learning method i.e. the Artificial Neural Networks. The aim of this comparison is to learn about the strengths and the weaknesses of these learning methods regarding different properties of their learning process. The comparison is made using two different comparison methods. In the first method the algorithm and the known behavioural model of these methods are analysed. Later, using this analysis, these methods are compared. In the second approach, a reference dataset that contains some of the most problematic features in the learning process is selected. Using the selected dataset the differences between two learning methods are numerically analysed and a comparison is made.

References

  1. Kabiri, P., 2002. Using Black Box and Machine Learning Approach in Controlling a Machine. In SMC'02, 2002 IEEE International Conference on Systems, Man and Cybernetics. IEEE.
  2. Kabiri, P., Sherkat, N. and Shih, V.,1999. Compensating for Robot Arm Positioning Inaccuracy in 3D-Space, In SPIE's International Symposium on Intelligent Systems in Design and Manufacturing II, SPIE.
  3. Kabiri, P., Sherkat, N. and Shih, V.,1998. Compensation for Robot Arm Flexibility Using Machine Intelligence, In SPIE's International Symposium on Intelligent Systems in Design and Manufacturing, SPIE.
  4. Mingzhong, L. and Fuli W., 1997. Adaptive Control of Black-Box Non-Linear Systems Using Recurrent Neural Networks, In Proceedings of the 36th IEEE Conference on Decision and Control, IEEE.
  5. McLauchlan, R. A., Ngamsom, P., Challoo, R. and Omar S. I., 1997, Three-Link Planar Robotic Arm Trajectory Planning and Its Adaptive Fuzzy Logic Control, In IEEE International Conference on Systems Man and Cybernetics, IEEE.
  6. Sima, J., 1996. Back-propagation is not efficient, In Neural Networks Journal.
  7. Chen, J. and Chao, L. M., 1986. Positioning Error Analysis for Robot Manipulators with All Rotary Joints, Proceedings of IEEE International Conference on Robotics and Automation, IEEE. Schaal, S. and Atkeson, C. G., 1994. Robot
Download


Paper Citation


in Harvard Style

Kabiri P. (2004). A COMPARISON BETWEEN THE PROPORTIONAL KEEN APPROXIMATOR AND THE NEURAL NETWORKS LEARNING METHODS . In Proceedings of the Sixth International Conference on Enterprise Information Systems - Volume 2: ICEIS, ISBN 972-8865-00-7, pages 159-164. DOI: 10.5220/0002599401590164


in Bibtex Style

@conference{iceis04,
author={Peyman Kabiri},
title={A COMPARISON BETWEEN THE PROPORTIONAL KEEN APPROXIMATOR AND THE NEURAL NETWORKS LEARNING METHODS},
booktitle={Proceedings of the Sixth International Conference on Enterprise Information Systems - Volume 2: ICEIS,},
year={2004},
pages={159-164},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0002599401590164},
isbn={972-8865-00-7},
}


in EndNote Style

TY - CONF
JO - Proceedings of the Sixth International Conference on Enterprise Information Systems - Volume 2: ICEIS,
TI - A COMPARISON BETWEEN THE PROPORTIONAL KEEN APPROXIMATOR AND THE NEURAL NETWORKS LEARNING METHODS
SN - 972-8865-00-7
AU - Kabiri P.
PY - 2004
SP - 159
EP - 164
DO - 10.5220/0002599401590164