A discretization technique converts continuous attribute values into discrete ones. Discretization is needed when classification algorithms require only discrete attributes. It is also useful to increase the speed and the accuracy of classification algorithms. This paper presents a dynamic discretization method, whose main characteristic is to detect interdependencies between all continuous attributes. Empirical evaluation on 12 datasets from the UCI repository shows that the proposed algorithm is a relatively effective method for discretization.
CITATION STYLE
Hwang, G. J., & Li, F. (2002). A dynamic method for discretization of continuous attributes. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 2412, pp. 506–511). Springer Verlag. https://doi.org/10.1007/3-540-45675-9_76
Mendeley helps you to discover research relevant for your work.