CRUISE Classification Tree (version 3.6.4)

© HyunJoong Kim 1997-2014

CRUISE is a statistical decision tree algorithm for classification (also called supervised learning) developed by Hyunjoong Kim (Yonsei University, Korea) and Wei-Yin Loh (University of Wisconsin-Madison, USA). It is a much-improved descendant of an older algorithm called FACT. CRUISE stands for Classification Rule with Unbiased Interaction Selection and Estimation.

CRUISE is unique among classification tree algorithms in possessing the following properties:

  • It splits each node into as many subnodes as the number of classes in the response variable
  • It has negligible bias in variable selection
  • It has several ways to deal with missing values
  • It can detect local interactions between pairs of predictor variables

See Table 1 for a feature comparison between CRUISE and other classification tree algorithms.

Documentation:

  1. Kim, H. and Loh, W.-Y. (2003), Classification trees with bivariate linear discriminant node models, Journal of Computational and Graphical Statistics, vol. 12, pp. 512-530. [This paper extends CRUISE to fit linear discriminant models in the leaf nodes.]
  2. Kim, H. and Loh, W.-Y. (2001), Classification trees with unbiased multiway splits, Journal of the American Statistical Association, vol. 96, pp. 589-604. [This is the principal reference for the CRUISE algorithm.]
  3. Loh, W.-Y. and Vanichsetakul, N. (1988), Tree-structured classification via generalized discriminant analysis (with discussion), Journal of the American Statistical Association, vol. 83, pp. 715-728. [This paper documents an earlier algorithm called FACT.]
  4. CRUISE manual in pdf format. The manual uses the example data and description files heartdat.txt and heartdsc.txt for illustration.

Compiled binaries: The following files may be freely distributed but not sold for profit.

  • Zipped Windows executable for Intel and compatibles --- download (download pkunzip.exe)
  • Gzipped Linux executable for Intel and compatibles --- download

Revision history: See the file history.txt

Tree diagrams: The CRUISE program can optionally produce LaTeX ( MikTeX) or allCLEAR source code for the tree diagrams. The LaTeX code requires the PSTricks package. Ghostscript and GSView are required to view the postscript files.

Related algorithms with unbiased splits:

  • QUEST: Classification trees with binary splits
  • GUIDE: Piecewise-linear least-squares, quantile, and Poisson regression trees
  • Application papers that use CRUISE, GUIDE, LOTUS, or QUEST: See file

    License:

    CRUISE is free software. You may use the Program without restriction. You may copy and distribute the Program in executable form provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice and disclaimer of warranty; and give any other recipients of the Program a copy of this license along with the Program.

    Disclaimer of Warranty:

    The copyright holder provides the Program "as is" without warranty of any kind, either expressed or implied, including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose. The entire risk as to the quality and performance of the Program is with you. Should the Program prove defective, you assume the cost of all necessary servicing, repair or correction. In no event will the copyright holder be liable to you for damages, including any general, special, incidental or consequential damages arising out of the use or inability to use the Program (including but not limited to loss of data or data being rendered inaccurate or losses sustained by you or third parties or a failure of the Program to operate with any other programs), even if such holder has been advised of the possibility of such damages.

    Return to Wei-Yin Loh's homepage for information on other decision tree algorithms.

    Last modified: September 13, 2014 by Wei-Yin Loh