Open Access
March 2013 Regression trees for longitudinal and multiresponse data
Wei-Yin Loh, Wei Zheng
Ann. Appl. Stat. 7(1): 495-522 (March 2013). DOI: 10.1214/12-AOAS596


Previous algorithms for constructing regression tree models for longitudinal and multiresponse data have mostly followed the CART approach. Consequently, they inherit the same selection biases and computational difficulties as CART. We propose an alternative, based on the GUIDE approach, that treats each longitudinal data series as a curve and uses chi-squared tests of the residual curve patterns to select a variable to split each node of the tree. Besides being unbiased, the method is applicable to data with fixed and random time points and with missing values in the response or predictor variables. Simulation results comparing its mean squared prediction error with that of MVPART are given, as well as examples comparing it with standard linear mixed effects and generalized estimating equation models. Conditions for asymptotic consistency of regression tree function estimates are also given.


Download Citation

Wei-Yin Loh. Wei Zheng. "Regression trees for longitudinal and multiresponse data." Ann. Appl. Stat. 7 (1) 495 - 522, March 2013.


Published: March 2013
First available in Project Euclid: 9 April 2013

zbMATH: 06171281
MathSciNet: MR3086428
Digital Object Identifier: 10.1214/12-AOAS596

Keywords: CART , decision tree , generalized estimating equation , linear mixed effects model , LOWESS , missing values , recursive partitioning , selection bias

Rights: Copyright © 2013 Institute of Mathematical Statistics

Vol.7 • No. 1 • March 2013
Back to Top