Hyper-rectangular space partitioning trees: A practical approach
(2005) Computational Statistics — Vol. 20, n° 1, p. 119-135 (2005)
Files
pdfdocument.pdf
Restricted Access - Adobe PDF
- 895.42 KB
Details
- Authors
- Abstract
- The process of computation of classification trees can be characterized as involving three basic choices: the type of splits considered in the growing process, the criterion to be optimized at each step of the process, and the way to get right-sized trees. Most implementations are ordinary binary trees, i.e. trees whose successive cuts are made by hyper-planes perpendicular to the axes. L. Devroye, L. Gyorfy and G. Lugosi (1996) define and consider the remarkable theoretical properties of a binary tree classifier whose prominent feature is the particular type of splits used in its construction: at a given node, partitioning is made by hyper-rectangles rather than hyper-planes. We propose an approximation of the solution for the complex optimization problem involved to allow insights on the practical advantages of those trees. Then we compare the performance of our algorithm with some leading algorithms for ordinary binary trees, namely CART and C4.5 as implemented in the Splus "tree" procedure and in SAS's Enterprise Miner respectively.
- Affiliations
Citations
De Macq, I., & Simar, L. (2005). Hyper-rectangular space partitioning trees: A practical approach. Computational Statistics, 20(1), 119-135. https://doi.org/10.1007/BF02736126 (Original work published 2005)