Multi-Interval Discretization of Continuous-Valued Attributes for Classification Learning Export

@inproceedings{citeulike:280566, abstract = {Since most real-world applications of classification learning involve continuous-valued attributes, properly addressing the discretization process is an important problem. This paper addresses the use of the entropy minimization heuristic for discretizing the range of a continuous-valued attribute into multiple intervals. We briefly present theoretical evidence for the appropriateness of this heuristic for use in the binary discretization algorithm used in ID3, C4, CART, and other learning algorithms. The results serve to justify extending the algorithm to derive multiple intervals. We formally derive a criterion based on the minimum description length principle for deciding the partitioning of intervals. We demonstrate via empirical evaluation on several real-world data sets that better decision trees are obtained using the new multi-interval algorithm.}, author = {Fayyad and Irani}, citeulike-article-id = {280566}, citeulike-linkout-0 = {http://www.cs.orst.edu/~bulatov/papers/fayyad-discretization.pdf}, journal = {Proceedings of the International Joint Conference on Uncertainty in AI}, keywords = {discretization}, pages = {1022--1027}, posted-at = {2005-08-13 04:28:07}, priority = {2}, title = {Multi-Interval Discretization of Continuous-Valued Attributes for Classification Learning}, url = {http://www.cs.orst.edu/~bulatov/papers/fayyad-discretization.pdf}, year = {1993} }

TY - CONF ID - citeulike:280566 L3 - citeulike-article-id:280566 N2 - Since most real-world applications of classification learning involve continuous-valued attributes, properly addressing the discretization process is an important problem. This paper addresses the use of the entropy minimization heuristic for discretizing the range of a continuous-valued attribute into multiple intervals. We briefly present theoretical evidence for the appropriateness of this heuristic for use in the binary discretization algorithm used in ID3, C4, CART, and other learning algorithms. The results serve to justify extending the algorithm to derive multiple intervals. We formally derive a criterion based on the minimum description length principle for deciding the partitioning of intervals. We demonstrate via empirical evaluation on several real-world data sets that better decision trees are obtained using the new multi-interval algorithm. JF - Proceedings of the International Joint Conference on Uncertainty in AI EP - 1027 TI - Multi-Interval Discretization of Continuous-Valued Attributes for Classification Learning SP - 1022 KW - discretization AU - Fayyad AU - Irani PY - 1993///Q334 .I571 UR - http://www.cs.orst.edu/~bulatov/papers/fayyad-discretization.pdf ER -