Abstract
In this paper, we propose a new design of a very simple data-driven binary classifier and conduct an empirical study of its performance. The data contain continuous and categorical variables . The classification system consists of highly interpretable fuzzy metarules. A new theorem is developed that guarantees that these metarules are equivalent to algebraic expressions . The algebraic expressions are obtained using the gene expression programming technique. The number of features in the modeled dataset does not affect the complexity of the metarules. The performance of the resulting metarules is comparable to that of the rules created by most of the popular machine learning methods . The newly introduced classifier (GPR) appears to be the simplest among the fuzzy rule-based classifiers. Its effectiveness was tested on 16 datasets and compared with 22 other classification algorithms . GPR turned out to be surprisingly good; i.e., it belongs to the group of the best classifiers when the quality criterion is the area under the ROC curve and the classification accuracy . The Scott-Knott analysis indicates that, in terms of performance, GPR is commensurate with the leading group of 3 algorithms, and the Wilcoxon test confirmed the statistical reliability of the obtained results. High interpretability is proved with examples of classification models .
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.