Abstract #301588

This is the preliminary program for the 2003 Joint Statistical Meetings in San Francisco, California. Currently included in this program is the "technical" program, schedule of invited, topic contributed, regular contributed and poster sessions; Continuing Education courses (August 2-5, 2003); and Committee and Business Meetings. This on-line program will be updated frequently to reflect the most current revisions.

To View the Program:
You may choose to view all activities of the program or just parts of it at any one time. All activities are arranged by date and time.

The views expressed here are those of the individual authors
and not necessarily those of the ASA or its board, officers, or staff.

Back to main JSM 2003 Program page
JSM 2003 Abstract #301588
Activity Number: 371
Type: Contributed
Date/Time: Wednesday, August 6, 2003 : 10:30 AM to 12:20 PM
Sponsor: Biometrics Section
Abstract - #301588
Title: Prediction of HIV Replication Capacity Using Trees and Forests
Author(s): Mark R. Segal*+
Companies: University of California, San Francisco
Address: Dept. of Biostatistics, San Francisco, CA, 94143-0560,
Keywords: HIV ; random forests ; regression trees ; sequence data ; viral fitness ; genotype
Abstract:

In the context of HIV disease, we investigate predicting a measure of viral fitness, replication capacity, based on amino acid sequence for positions 4 to 99 of protease and 38 to 223 of reverse transcriptase. The sequence data for these 282 positions is obtained by sequencing viral samples from a cohort study of 336 patients. A forefront statistical concern is handling this large number of multilevel (potentially 20 amino acids per postion), unordered categorical covariates. The need to further accommodate between-position interactions makes conventional linear predictor based methods utilizing indicators problematic. We have proposed use of tree-based approaches in such settings. Applying such methods here yields interesting results with respect to the specific positions deemed consequential. However, tree techniques have limitations, especially pertaining to predictive ability. Random forests (Breiman 2001) have been shown to be "A+" predictors that, importantly here, yield mechanistic insight. So we investigate the performance of random forests and provide a critique as to their appropriateness for this class of problems.

  • The address information is for the authors that have a + after their name.
  • Authors who are presenting talks have a * after their name.

Back to the full JSM 2003 program

JSM 2003 For information, contact meetings@amstat.org or phone (703) 684-1221. If you have questions about the Continuing Education program, please contact the Education Department.
Revised March 2003