Abstract
The paper presents three algorithms of instance selection for regression problems, which extend the capabilities of the CNN, ENN and CA algorithms used for classification tasks. Various combinations of the algorithms are experimentally evaluated as data preprocessing for regression tree induction. The influence of the instance selection algorithms and their parameters on the accuracy and rules produced by regression trees is evaluated and compared to the results obtained with tree pruning. © 2013 Springer-Verlag.
Cite
CITATION STYLE
Kordos, M., Białka, S., & Blachnik, M. (2013). Instance selection in logical rule extraction for regression problems. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7895 LNAI, pp. 167–175). https://doi.org/10.1007/978-3-642-38610-7_16
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
