Applications of inductive learning techniques: a survey in the Netherlands

Article type: Research Article

Authors: Verdenius, Floor | van Someren, Maarten W.

Affiliations: ATO‐DLO, Postbus 17, 6700 AA Wageningen, Netherlands E‐mail: F.Verdenius@ ato.dlo.nl | Department of Social Science Informatics, Faculty of Psychology, University of Amsterdam, Roeterstraat 15, 1018 WB Amsterdam, Netherlands E‐mail: maarten@ swi.psy.uva.nl

Abstract: Interest for machine learning techniques has increased over the last decade. In spite of this the application practice of these techniques has never been systematically analysed. This paper analyses the practical application of Inductive Learning Techniques in the Netherlands by means of a survey. Results of this survey are assessed in terms of introduction of (information) technological innovation. The application practice for these techniques finds itself in an initial stage. Current practice is dominated by technical issues and there is little attention for methodological issues associated with analysis of the problem and data collection. In the paper we propose a four level model for describing the methodological aspects of ILT application. The current practice of application concentrates on the two lowermost layers. Tools for developing applications concentrate mainly on the lowest, technical level.

Journal: AI Communications, vol. 10, no. 1, pp. 3-20, 1997