Classification Trees

  • Classification from Ordered Predictors
    Complete implementation of the discriminant-based linear combination splits method of QUEST (Quick, Unbiased, Efficient Statistical Trees) for classification from ordered predictors (see also the Classification Trees for Categorical and Ordered Predictors facilities). QUEST is a classification tree algorithm developed by Loh and Shih (1997) that employs a modification of recursive quadratic discriminant analysis and includes a number of innovative features for improving the reliability and efficiency of the classification trees that it computes. The program allows you to specify continuous (ordered) predictors, and learning and testing samples; if a testing sample is specified, predicted classifications can be computed based on the final classification solution. See also the General C & RT methods and General CHAID methods for tree classifiers with large data sets (with many classes) and ANCOVA-like predictor designs.
  • Classification from Categorical and Ordered Predictors
    Complete implementation of the discriminant-based univariate splits method of QUEST (Quick, Unbiased, Efficient Statistical Trees) for classification from categorical and ordered predictors. QUEST is a classification tree algorithm developed by Loh and Shih (1997) that employs a modification of recursive quadratic discriminant analysis and includes a number of innovative features for improving the reliability and efficiency of the classification trees that it computes. The program allows you to specify categorical and continuous (ordered) predictors, and learning and testing samples; if a testing sample is specified, predicted classifications can be computed based on the final classification solution. See also the General C&RT methods and General CHAID methods for tree classifiers with large data sets (with many classes) and ANCOVA-like predictor designs.
  • Exhaustive (C and RT) Search for Univariate Splits
    Implementation of C and RT (Classification and Regression Trees) style search for univariate splits, based on the QUEST (Quick, Unbiased, Efficient Statistical Trees) algorithms for classification. See also the designated C & RT (and CHAID) methods available as part of Data Miner for more comprehensive implementations of these techniques.