Targeted use of growth mixture modeling

A learning perspective

Booil Jo, Robert L Findling, Chen Pin Wang, Trevor J. Hastie, Eric A. Youngstrom, L. Eugene Arnold, Mary A. Fristad, Sarah Mccue Horwitz

Research output: Contribution to journalArticle

Abstract

From the statistical learning perspective, this paper shows a new direction for the use of growth mixture modeling (GMM), a method of identifying latent subpopulations that manifest heterogeneous outcome trajectories. In the proposed approach, we utilize the benefits of the conventional use of GMM for the purpose of generating potential candidate models based on empirical model fitting, which can be viewed as unsupervised learning. We then evaluate candidate GMM models on the basis of a direct measure of success; how well the trajectory types are predicted by clinically and demographically relevant baseline features, which can be viewed as supervised learning. We examine the proposed approach focusing on a particular utility of latent trajectory classes, as outcomes that can be used as valid prediction targets in clinical prognostic models. Our approach is illustrated using data from the Longitudinal Assessment of Manic Symptoms study.

Original languageEnglish (US)
JournalStatistics in Medicine
DOIs
StateAccepted/In press - 2016

Fingerprint

Mixture Modeling
Learning
Trajectory
Growth
Statistical Learning
Symptom Assessment
Empirical Model
Unsupervised Learning
Model Fitting
Supervised Learning
Baseline
Valid
Model-based
Target
Evaluate
Prediction
Model

Keywords

  • Early prediction
  • Growth mixture modeling
  • Latent trajectory class
  • Sensitivity
  • Specificity
  • Supervised learning
  • Unsupervised learning

ASJC Scopus subject areas

  • Epidemiology
  • Statistics and Probability

Cite this

Jo, B., Findling, R. L., Wang, C. P., Hastie, T. J., Youngstrom, E. A., Arnold, L. E., ... Horwitz, S. M. (Accepted/In press). Targeted use of growth mixture modeling: A learning perspective. Statistics in Medicine. https://doi.org/10.1002/sim.7152

Targeted use of growth mixture modeling : A learning perspective. / Jo, Booil; Findling, Robert L; Wang, Chen Pin; Hastie, Trevor J.; Youngstrom, Eric A.; Arnold, L. Eugene; Fristad, Mary A.; Horwitz, Sarah Mccue.

In: Statistics in Medicine, 2016.

Research output: Contribution to journalArticle

Jo, B, Findling, RL, Wang, CP, Hastie, TJ, Youngstrom, EA, Arnold, LE, Fristad, MA & Horwitz, SM 2016, 'Targeted use of growth mixture modeling: A learning perspective', Statistics in Medicine. https://doi.org/10.1002/sim.7152
Jo, Booil ; Findling, Robert L ; Wang, Chen Pin ; Hastie, Trevor J. ; Youngstrom, Eric A. ; Arnold, L. Eugene ; Fristad, Mary A. ; Horwitz, Sarah Mccue. / Targeted use of growth mixture modeling : A learning perspective. In: Statistics in Medicine. 2016.
@article{ed0074c3f0a44254a9061a34ffa72397,
title = "Targeted use of growth mixture modeling: A learning perspective",
abstract = "From the statistical learning perspective, this paper shows a new direction for the use of growth mixture modeling (GMM), a method of identifying latent subpopulations that manifest heterogeneous outcome trajectories. In the proposed approach, we utilize the benefits of the conventional use of GMM for the purpose of generating potential candidate models based on empirical model fitting, which can be viewed as unsupervised learning. We then evaluate candidate GMM models on the basis of a direct measure of success; how well the trajectory types are predicted by clinically and demographically relevant baseline features, which can be viewed as supervised learning. We examine the proposed approach focusing on a particular utility of latent trajectory classes, as outcomes that can be used as valid prediction targets in clinical prognostic models. Our approach is illustrated using data from the Longitudinal Assessment of Manic Symptoms study.",
keywords = "Early prediction, Growth mixture modeling, Latent trajectory class, Sensitivity, Specificity, Supervised learning, Unsupervised learning",
author = "Booil Jo and Findling, {Robert L} and Wang, {Chen Pin} and Hastie, {Trevor J.} and Youngstrom, {Eric A.} and Arnold, {L. Eugene} and Fristad, {Mary A.} and Horwitz, {Sarah Mccue}",
year = "2016",
doi = "10.1002/sim.7152",
language = "English (US)",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",

}

TY - JOUR

T1 - Targeted use of growth mixture modeling

T2 - A learning perspective

AU - Jo, Booil

AU - Findling, Robert L

AU - Wang, Chen Pin

AU - Hastie, Trevor J.

AU - Youngstrom, Eric A.

AU - Arnold, L. Eugene

AU - Fristad, Mary A.

AU - Horwitz, Sarah Mccue

PY - 2016

Y1 - 2016

N2 - From the statistical learning perspective, this paper shows a new direction for the use of growth mixture modeling (GMM), a method of identifying latent subpopulations that manifest heterogeneous outcome trajectories. In the proposed approach, we utilize the benefits of the conventional use of GMM for the purpose of generating potential candidate models based on empirical model fitting, which can be viewed as unsupervised learning. We then evaluate candidate GMM models on the basis of a direct measure of success; how well the trajectory types are predicted by clinically and demographically relevant baseline features, which can be viewed as supervised learning. We examine the proposed approach focusing on a particular utility of latent trajectory classes, as outcomes that can be used as valid prediction targets in clinical prognostic models. Our approach is illustrated using data from the Longitudinal Assessment of Manic Symptoms study.

AB - From the statistical learning perspective, this paper shows a new direction for the use of growth mixture modeling (GMM), a method of identifying latent subpopulations that manifest heterogeneous outcome trajectories. In the proposed approach, we utilize the benefits of the conventional use of GMM for the purpose of generating potential candidate models based on empirical model fitting, which can be viewed as unsupervised learning. We then evaluate candidate GMM models on the basis of a direct measure of success; how well the trajectory types are predicted by clinically and demographically relevant baseline features, which can be viewed as supervised learning. We examine the proposed approach focusing on a particular utility of latent trajectory classes, as outcomes that can be used as valid prediction targets in clinical prognostic models. Our approach is illustrated using data from the Longitudinal Assessment of Manic Symptoms study.

KW - Early prediction

KW - Growth mixture modeling

KW - Latent trajectory class

KW - Sensitivity

KW - Specificity

KW - Supervised learning

KW - Unsupervised learning

UR - http://www.scopus.com/inward/record.url?scp=84996588097&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84996588097&partnerID=8YFLogxK

U2 - 10.1002/sim.7152

DO - 10.1002/sim.7152

M3 - Article

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

ER -