Abstract
Experimentation has an important role in determining the capacities and restrictions of machine learning (ML) systems. In this paper we present the definition of some sensitivity and evaluation criteria which can be used to perform an evaluation of learning systems. Moreover, in order to overcome some of the limitations of real data sets, we introduce the specification of a parametrable generator of artificial learning sets which allows us to make easily complete experiments to discover some empirical rules of behavior for ML algorithms. Finally, we give some results obtained with different algorithms, showing that artificial data bases approach is an interesting direction to explore.
This work is partially supported by CEC through the ESPRIT-2 contract MLT 2154 ("Machine Learning Toolbox") and also by MRT through PRC-IA.
Chapter PDF
Keywords
References
Bisson, Laublet 1989, Méthode d'évaluation des systèmes d'apprentissage, Rapport de fin d'étude, ONERA-GIA, August 1989.
Bisson, Laublet 1990, Theoritical and empirical study of learning systems, Cognitiva 1990 Madrid, 1990.
Bisson 1991, Learning of Rule Systems in a First Order Representation, LRI Internal report 1991
Bonelli 1990, A.PARODI, S.SEN, S.WILSON, NewBOOLE: A fast GBML system, 7th ICML, Austin, 1990.
Boswell 1990 a, Manual for CN2, The Turing Institute, January 1990.
Boswell 1990 b, Manual for NewID version 2.1, The Turing Institute, January 1990.
Cannat 1988, Machine Learning Applied to Air Trafic Control, Proc. Human-Machine Interaction Artificial Intelligence Aeronautics & Space, pp.265–274, Toulouse, 1988.
Elomaa 1989, N.HOLSTI, An experimental comparison of inducing decision trees and decision lists in noisy domains, EWSL Montpellier, 1989.
Ganascia 1987, AGAPE et CHARADE: deux techniques d'apprentissage symbolique appliquées à la construction de bases de connaissances, Thèse d'état soutenue 4/87,Université Paris Sud.
Ganascia, Helft 1988, Evaluation des Systèmes d'Apprentissage, 3éme Journées Françaises de l'Apprentissage 1988 Cassis, p 3–20.
Kibler, Langley 1988, Machine Learning as an Experimental Science, EWSL 1988, Glasgow, pp 81–92.
Manago, Kodratoff 1987, Noise and Knowledge Acquisition, Proceeding of 10th IJCAI, Los Altos, 1987
Martin 1989, Reducing Redundant Learning, in Proceedings of the 6o International Workshop on Machine Learning, Ithaca 1989, p 396–99.
Minton 1985, Selectively Generalizing Plans for Problem Solving, IJCAI 85 p596–599.
MLT 1990, Laboratoire de Marcoussis, British Aerospace, Université de Paris-Sud, Evaluation Methods for Attribute-Based Algorithms, Deliverable 7.2, Projet ESPRIT 2154, Aout 1990.
Quinlan 1986, The effect of noise on concept learning, Machine learning 2 an artificial intelligence approach, TIOGA 1986.
Quinlan 1988, Symplifying decision trees, in B.Gaines, J.Boose Knowledge Acquisition for Knowledge-Based Systems, Academic Press 1988.
Rendell 1989, H. CHO, R. SESHU Improving the Design of Similarity-Based Rule-Learning Systems, International Journal of Expert System, pp 97–133, volume 2, Number 1, 1989.
Rumelhart 1986, G.E. HINTON, R.J. WILLIAMS, Learning internal representation by error propagation. Parallel distributed processing: Explorations in the microstructures of cognition, MIT Press, vol 1, pp 318–362, 1986.
Shavlik 1989, An Empirical Analysis of EBL Approaches for Learning Plan Schemata, in Proceedings of the 6o International Workshop on Machine Learning, Ithaca 1989, p183–87.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1991 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Lounis, H., Bisson, G. (1991). Evaluation of learning systems : An artificial data-based approach. In: Kodratoff, Y. (eds) Machine Learning — EWSL-91. EWSL 1991. Lecture Notes in Computer Science, vol 482. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0017038
Download citation
DOI: https://doi.org/10.1007/BFb0017038
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-53816-5
Online ISBN: 978-3-540-46308-5
eBook Packages: Springer Book Archive