This is the program for the 2010 Joint Statistical Meetings in Vancouver, British Columbia.

Abstract Details

Activity Number: 479
Type: Contributed
Date/Time: Wednesday, August 4, 2010 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistical Learning and Data Mining
Abstract - #308232
Title: PartDSA: The Partitioning Deletion/Substitution/Addition Algorithm for Creating Survival Risk Groups
Author(s): Karen Lostritto*+ and Rob Strawderman and Annette Molinaro+
Companies: Yale University and Cornell University and Yale University
Address: , New Haven, CT, , , New Haven, CT, ,
Keywords: survival outcomes ; brier score ; IPCW ; partDSA ; recursive partitioning ; prediction
Abstract:

The ability to accurately stratify patients into risk groups based on a potentially censored outcome is a crucial step towards improving the effectiveness of clinical decision making and subsequent treatment. We have previously compared partDSA to other statistical learning methods in the setting with always observed categorical and continuous outcomes. In such scenarios, partDSA created the most parsimonious models based on the correct predictor variables, resulting in the lowest out-of-sample prediction error. Here we extend partDSA to censored outcomes through the use of modified loss functions, specifically the IPCW squared error and Brier loss functions. We show both in simulation studies and data analysis that partDSA surpasses other survival outcome methods by substantially decreasing prediction error and accurately selecting relevant predictor variables.


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 2010 program




2010 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.