Literature DB >> 36261805

Estimating the optimal linear combination of predictors using spherically constrained optimization.

Priyam Das1, Debsurya De2, Raju Maiti3, Mona Kamal4, Katherine A Hutcheson5, Clifton D Fuller4, Bibhas Chakraborty3,6,7, Christine B Peterson8.   

Abstract

BACKGROUND: In the context of a binary classification problem, the optimal linear combination of continuous predictors can be estimated by maximizing the area under the receiver operating characteristic curve. For ordinal responses, the optimal predictor combination can similarly be obtained by maximization of the hypervolume under the manifold (HUM). Since the empirical HUM is discontinuous, non-differentiable, and possibly multi-modal, solving this maximization problem requires a global optimization technique. Estimation of the optimal coefficient vector using existing global optimization techniques is computationally expensive, becoming prohibitive as the number of predictors and the number of outcome categories increases.
RESULTS: We propose an efficient derivative-free black-box optimization technique based on pattern search to solve this problem, which we refer to as Spherically Constrained Optimization Routine (SCOR). Through extensive simulation studies, we demonstrate that the proposed method achieves better performance than existing methods including the step-down algorithm. Finally, we illustrate the proposed method to predict the severity of swallowing difficulty after radiation therapy for oropharyngeal cancer based on radiation dose to various structures in the head and neck.
CONCLUSIONS: Our proposed method addresses an important challenge in combining multiple biomarkers to predict an ordinal outcome. This problem is particularly relevant to medical research, where it may be of interest to diagnose a disease with various stages of progression or a toxicity with multiple grades of severity. We provide the implementation of our proposed SCOR method as an R package, available online at https://CRAN.R-project.org/package=SCOR .
© 2022. The Author(s).

Entities:  

Keywords:  Area under the curve; Classification; Global optimization; Hypervolume under the manifold; Pattern search; ROC curve

Mesh:

Substances:

Year:  2022        PMID: 36261805      PMCID: PMC9583504          DOI: 10.1186/s12859-022-04953-y

Source DB:  PubMed          Journal:  BMC Bioinformatics        ISSN: 1471-2105            Impact factor:   3.307


  24 in total

1.  Combining diagnostic test results to increase accuracy.

Authors:  M S Pepe; M L Thompson
Journal:  Biostatistics       Date:  2000-06       Impact factor: 5.899

2.  Optimal linear combination of biomarkers for multi-category diagnosis.

Authors:  Man-Jen Hsu; Yi-Hau Chen
Journal:  Stat Med       Date:  2015-08-06       Impact factor: 2.373

3.  Ordered multiple-class ROC analysis with continuous measurements.

Authors:  Christos T Nakas; Constantin T Yiannoutsos
Journal:  Stat Med       Date:  2004-11-30       Impact factor: 2.373

4.  Optimization by simulated annealing.

Authors:  S Kirkpatrick; C D Gelatt; M P Vecchi
Journal:  Science       Date:  1983-05-13       Impact factor: 47.728

5.  Index for rating diagnostic tests.

Authors:  W J YOUDEN
Journal:  Cancer       Date:  1950-01       Impact factor: 6.860

6.  Three-way ROCs.

Authors:  D Mossman
Journal:  Med Decis Making       Date:  1999 Jan-Mar       Impact factor: 2.583

7.  Multiple-Event Forced-Choice Tasks in the Theory of Signal Detectability

Authors: 
Journal:  J Math Psychol       Date:  1996-09       Impact factor: 2.223

8.  A min-max combination of biomarkers to improve diagnostic accuracy.

Authors:  Chunling Liu; Aiyi Liu; Susan Halabi
Journal:  Stat Med       Date:  2011-04-07       Impact factor: 2.373

9.  Linear combinations of biomarkers to improve diagnostic accuracy with three ordinal diagnostic categories.

Authors:  Le Kang; Chengjie Xiong; Paul Crane; Lili Tian
Journal:  Stat Med       Date:  2012-08-03       Impact factor: 2.373

10.  Optimization by Adaptive Stochastic Descent.

Authors:  Cliff C Kerr; Salvador Dura-Bernal; Tomasz G Smolinski; George L Chadderdon; David P Wilson
Journal:  PLoS One       Date:  2018-03-16       Impact factor: 3.240

View more

北京卡尤迪生物科技股份有限公司 © 2022-2023.