A weighted random forests approach to improve predictive performance

Research output: Contribution to journalArticle

25 Citations (Scopus)

Abstract

Identifying genetic variants associated with complex disease in high-dimensional data is a challenging problem, and complicated etiologies such as gene-gene interactions are often ignored in analyses. The data-mining method random forests (RF) can handle high dimensions; however, in high-dimensional data, RF is not an effective filter for identifying risk factors associated with the disease trait via complex genetic models such as gene-gene interactions without strong marginal components. In this article we propose an extension called weighted random forests (wRF), which incorporates tree-level weights to emphasize more accurate trees in prediction and calculation of variable importance. We demonstrate through simulation and application to data from a genetic study of addiction that wRF can outperform RF in high-dimensional data, although the improvements are modest and limited to situations with effect sizes that are larger than what is realistic in genetics of complex disease. Thus, the current implementation of wRF is unlikely to improve detection of relevant predictors in high-dimensional genetic data, but may be applicable in other situations where larger effect sizes are anticipated.

Original languageEnglish (US)
Pages (from-to)496-505
Number of pages10
JournalStatistical Analysis and Data Mining
Volume6
Issue number6
DOIs
StatePublished - Dec 2013

Fingerprint

Random Forest
Genes
High-dimensional Data
Gene
Effect Size
Trees (mathematics)
Data mining
Risk Factors
Interaction
Higher Dimensions
Predictors
Data Mining
High-dimensional
Filter
Prediction
Demonstrate
Simulation

Keywords

  • Gene-gene interactions
  • Genetic data
  • Genome-wide association
  • High-dimensional data
  • Random forests
  • Weighting

ASJC Scopus subject areas

  • Information Systems
  • Computer Science Applications
  • Analysis

Cite this

@article{f3b3054071ec44a69d38156c2074a9df,
title = "A weighted random forests approach to improve predictive performance",
abstract = "Identifying genetic variants associated with complex disease in high-dimensional data is a challenging problem, and complicated etiologies such as gene-gene interactions are often ignored in analyses. The data-mining method random forests (RF) can handle high dimensions; however, in high-dimensional data, RF is not an effective filter for identifying risk factors associated with the disease trait via complex genetic models such as gene-gene interactions without strong marginal components. In this article we propose an extension called weighted random forests (wRF), which incorporates tree-level weights to emphasize more accurate trees in prediction and calculation of variable importance. We demonstrate through simulation and application to data from a genetic study of addiction that wRF can outperform RF in high-dimensional data, although the improvements are modest and limited to situations with effect sizes that are larger than what is realistic in genetics of complex disease. Thus, the current implementation of wRF is unlikely to improve detection of relevant predictors in high-dimensional genetic data, but may be applicable in other situations where larger effect sizes are anticipated.",
keywords = "Gene-gene interactions, Genetic data, Genome-wide association, High-dimensional data, Random forests, Weighting",
author = "Winham, {Stacey J} and Robert Freimuth and Biernacka, {Joanna M}",
year = "2013",
month = "12",
doi = "10.1002/sam.11196",
language = "English (US)",
volume = "6",
pages = "496--505",
journal = "Statistical Analysis and Data Mining",
issn = "1932-1872",
publisher = "John Wiley and Sons Inc.",
number = "6",

}

TY - JOUR

T1 - A weighted random forests approach to improve predictive performance

AU - Winham, Stacey J

AU - Freimuth, Robert

AU - Biernacka, Joanna M

PY - 2013/12

Y1 - 2013/12

N2 - Identifying genetic variants associated with complex disease in high-dimensional data is a challenging problem, and complicated etiologies such as gene-gene interactions are often ignored in analyses. The data-mining method random forests (RF) can handle high dimensions; however, in high-dimensional data, RF is not an effective filter for identifying risk factors associated with the disease trait via complex genetic models such as gene-gene interactions without strong marginal components. In this article we propose an extension called weighted random forests (wRF), which incorporates tree-level weights to emphasize more accurate trees in prediction and calculation of variable importance. We demonstrate through simulation and application to data from a genetic study of addiction that wRF can outperform RF in high-dimensional data, although the improvements are modest and limited to situations with effect sizes that are larger than what is realistic in genetics of complex disease. Thus, the current implementation of wRF is unlikely to improve detection of relevant predictors in high-dimensional genetic data, but may be applicable in other situations where larger effect sizes are anticipated.

AB - Identifying genetic variants associated with complex disease in high-dimensional data is a challenging problem, and complicated etiologies such as gene-gene interactions are often ignored in analyses. The data-mining method random forests (RF) can handle high dimensions; however, in high-dimensional data, RF is not an effective filter for identifying risk factors associated with the disease trait via complex genetic models such as gene-gene interactions without strong marginal components. In this article we propose an extension called weighted random forests (wRF), which incorporates tree-level weights to emphasize more accurate trees in prediction and calculation of variable importance. We demonstrate through simulation and application to data from a genetic study of addiction that wRF can outperform RF in high-dimensional data, although the improvements are modest and limited to situations with effect sizes that are larger than what is realistic in genetics of complex disease. Thus, the current implementation of wRF is unlikely to improve detection of relevant predictors in high-dimensional genetic data, but may be applicable in other situations where larger effect sizes are anticipated.

KW - Gene-gene interactions

KW - Genetic data

KW - Genome-wide association

KW - High-dimensional data

KW - Random forests

KW - Weighting

UR - http://www.scopus.com/inward/record.url?scp=84890157266&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84890157266&partnerID=8YFLogxK

U2 - 10.1002/sam.11196

DO - 10.1002/sam.11196

M3 - Article

VL - 6

SP - 496

EP - 505

JO - Statistical Analysis and Data Mining

JF - Statistical Analysis and Data Mining

SN - 1932-1872

IS - 6

ER -