Constructive Induction methods aim to solve the problem of learning hard concepts despite complex interaction in data. We propose a new Constructive Induction method based on Genetic Algorithms with a non-algebraic representation of features. The advantage of our method to some other similar methods is that it constructs and evaluates a combination of features. Evaluating constructed features together, instead of considering them one by one, is essential when number of interacting attributes is high and there are more than one interaction in concept. Our experiments show the effectiveness of this method to learn such concepts. © Springer-Verlag Berlin Heidelberg 2004.
CITATION STYLE
Shafti, L. S., & Pérez, E. (2004). Machine learning by multi-feature extraction using Genetic Algorithms. In Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science) (Vol. 3315, pp. 246–255). Springer Verlag. https://doi.org/10.1007/978-3-540-30498-2_25
Mendeley helps you to discover research relevant for your work.