OPTIMAL PARTITIONING FOR CLASSIFICATION AND REGRESSION TREES

被引:148
|
作者
CHOU, PA [1 ]
机构
[1] STANFORD UNIV, DEPT ELECT ENGN, STANFORD, CA 94305 USA
关键词
CLUSTERING; DECISION TREES; INFORMATION DIVERGENCE; TEXT-TO-SPEECH;
D O I
10.1109/34.88569
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In designing a decision tree for classification or regression, one selects at each node a feature to be tested, and partitions the range of that feature into a small number of bins, each bin corresponding to a child of the node. When the feature's range is discrete with N unordered outcomes, the optimal partition, that is, the partition minimizing an expected loss, is usually found by an exhaustive search through all possible partitions. Since the number of possible partitions grows exponentially in N, this approach is impractical when N is larger than about 10 or 20. In this paper, we present an iterative algorithm that finds a locally optimal partition for an arbitrary loss function, in time linear in N for each iteration. The algorithm is a K-means like clustering algorithm that uses as its distance measure a generalization of Kullback's information divergence. Moreover, we prove that the globally optimal partition must satisfy a nearest neighbor condition using divergence as the distance measure. These results generalize similar results of Breiman et al. to an arbitrary number of classes or regression variables and to an arbitrary number of bins. We also provide experimental results on a text-to-speech example, and we suggest additional applications of the algorithm, including the design of variable combinations, surrogate splits, composite nodes, and decision graphs.
引用
收藏
页码:340 / 354
页数:15
相关论文
共 50 条
  • [1] Globally optimal fuzzy decision trees for classification and regression
    Suárez, A
    Lutsko, JF
    IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1999, 21 (12) : 1297 - 1311
  • [2] Classification and regression trees
    Martin Krzywinski
    Naomi Altman
    Nature Methods, 2017, 14 : 757 - 758
  • [3] Classification and regression trees
    Loh, Wei-Yin
    WILEY INTERDISCIPLINARY REVIEWS-DATA MINING AND KNOWLEDGE DISCOVERY, 2011, 1 (01) : 14 - 23
  • [4] Classification and regression trees
    Speybroeck, N.
    INTERNATIONAL JOURNAL OF PUBLIC HEALTH, 2012, 57 (01) : 243 - 246
  • [5] Classification and regression trees
    Krzywinski, Martin
    Altman, Naomi
    NATURE METHODS, 2017, 14 (08) : 755 - 756
  • [6] ectree: Evolutionary Learning of Globally Optimal Classification and Regression Trees in R
    Grubinger, Thomas
    Zeileis, Achim
    Pfeiffer, Karl-Peter
    JOURNAL OF STATISTICAL SOFTWARE, 2014, 61 (01): : 1 - 29
  • [7] Optimal classification trees
    Dimitris Bertsimas
    Jack Dunn
    Machine Learning, 2017, 106 : 1039 - 1082
  • [8] Optimal classification trees
    Bertsimas, Dimitris
    Dunn, Jack
    MACHINE LEARNING, 2017, 106 (07) : 1039 - 1082
  • [9] Decision trees with optimal joint partitioning
    Zighed, DA
    Ritschard, G
    Erray, W
    Scuturici, VM
    INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS, 2005, 20 (07) : 693 - 718
  • [10] On sparse optimal regression trees
    Blanquero, Rafael
    Carrizosa, Emilio
    Molero-Rio, Cristina
    Morales, Dolores Romero
    EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2022, 299 (03) : 1045 - 1054