DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Regression models using shapes of functions as predictors

Abstract

Functional variables are often used as predictors in regression problems. A commonly used parametric approach, called scalar-on-function regression, uses the $$\mathbb L^2$$ inner product to map functional predictors into scalar responses. This method can perform poorly when predictor functions contain undesired phase variability, causing phases to have disproportionately large influence on the response variable. One past solution has been to perform phase–amplitude separation (as a pre-processing step) and then use only the amplitudes in the regression model. In this paper, we propose a more integrated approach, termed elastic functional regression model (EFRM), where phase-separation is performed inside the regression model, rather than as a pre-processing step. This approach generalizes the notion of phase in functional data, and is based on the norm-preserving time warping of predictors. Due to its invariance properties, this representation provides robustness to predictor phase variability and results in improved predictions of the response variable over traditional models. We demonstrate this framework using a number of datasets involving gait signals, NMR data, and stock market prices.

Authors:
 [1]; ORCiD logo [2];  [3];  [3]
  1. RIKEN Center for Biosystems Dynamics Research (BDR), Kobe (Japan)
  2. Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
  3. Florida State Univ., Tallahassee, FL (United States)
Publication Date:
Research Org.:
Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
Sponsoring Org.:
USDOE National Nuclear Security Administration (NNSA); USDOE Laboratory Directed Research and Development (LDRD) Program; National Science Foundation (NSF)
OSTI Identifier:
1634794
Alternate Identifier(s):
OSTI ID: 1776473
Report Number(s):
SAND-2020-5389J
Journal ID: ISSN 0167-9473; 686289
Grant/Contract Number:  
AC04-94AL85000; NA0003525; 1621787; 1617397
Resource Type:
Accepted Manuscript
Journal Name:
Computational Statistics and Data Analysis (Print)
Additional Journal Information:
Journal Name: Computational Statistics and Data Analysis (Print); Journal Volume: 151; Journal Issue: C; Journal ID: ISSN 0167-9473
Publisher:
Elsevier
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICS AND COMPUTING; functional data analysis; scalar-on-function regression; functional single-index model; function alignment; SRVF

Citation Formats

Ahn, Kyungmin, Tucker, J. Derek, Wu, Wei, and Srivastava, Anuj. Regression models using shapes of functions as predictors. United States: N. p., 2020. Web. doi:10.1016/j.csda.2020.107017.
Ahn, Kyungmin, Tucker, J. Derek, Wu, Wei, & Srivastava, Anuj. Regression models using shapes of functions as predictors. United States. https://doi.org/10.1016/j.csda.2020.107017
Ahn, Kyungmin, Tucker, J. Derek, Wu, Wei, and Srivastava, Anuj. Sat . "Regression models using shapes of functions as predictors". United States. https://doi.org/10.1016/j.csda.2020.107017. https://www.osti.gov/servlets/purl/1634794.
@article{osti_1634794,
title = {Regression models using shapes of functions as predictors},
author = {Ahn, Kyungmin and Tucker, J. Derek and Wu, Wei and Srivastava, Anuj},
abstractNote = {Functional variables are often used as predictors in regression problems. A commonly used parametric approach, called scalar-on-function regression, uses the $\mathbb L^2$ inner product to map functional predictors into scalar responses. This method can perform poorly when predictor functions contain undesired phase variability, causing phases to have disproportionately large influence on the response variable. One past solution has been to perform phase–amplitude separation (as a pre-processing step) and then use only the amplitudes in the regression model. In this paper, we propose a more integrated approach, termed elastic functional regression model (EFRM), where phase-separation is performed inside the regression model, rather than as a pre-processing step. This approach generalizes the notion of phase in functional data, and is based on the norm-preserving time warping of predictors. Due to its invariance properties, this representation provides robustness to predictor phase variability and results in improved predictions of the response variable over traditional models. We demonstrate this framework using a number of datasets involving gait signals, NMR data, and stock market prices.},
doi = {10.1016/j.csda.2020.107017},
journal = {Computational Statistics and Data Analysis (Print)},
number = C,
volume = 151,
place = {United States},
year = {Sat May 30 00:00:00 EDT 2020},
month = {Sat May 30 00:00:00 EDT 2020}
}

Journal Article:

Citation Metrics:
Cited by: 2 works
Citation information provided by
Web of Science

Save / Share:

Works referenced in this record:

Cross-validated estimations in the single-functional index model
journal, December 2008


Prediction in functional linear regression
journal, October 2006


Functional linear model
journal, October 1999


Wavelet-based scalar-on-function finite mixture regression models
journal, January 2016


Multivariate calibration with single-index signal regression
journal, April 2009


Statistical Computing in Functional Data Analysis: The R Package fda.usc
journal, January 2012

  • Febrero-Bande, Manuel; Fuente, Manuel Oviedo de la
  • Journal of Statistical Software, Vol. 51, Issue 4
  • DOI: 10.18637/jss.v051.i04

Treadmill walking as an external pacemaker to improve gait rhythm and stability in Parkinson's disease
journal, January 2005

  • Frenkel-Toledo, Silvi; Giladi, Nir; Peretz, Chava
  • Movement Disorders, Vol. 20, Issue 9
  • DOI: 10.1002/mds.20507

Penalized scalar-on-functions regression with interaction term
journal, January 2015


A Goodness-of-Fit Test for the Functional Linear Model with Scalar Response
journal, June 2014

  • García-Portugués, Eduardo; González-Manteiga, Wenceslao; Febrero-Bande, Manuel
  • Journal of Computational and Graphical Statistics, Vol. 23, Issue 3
  • DOI: 10.1080/10618600.2013.812519

Longitudinal scalar-on-functions regression with application to tractography data
journal, January 2013


PhysioBank, PhysioToolkit, and PhysioNet: Components of a New Research Resource for Complex Physiologic Signals
journal, June 2000


Estimator selection and combination in scalar-on-function regression
journal, February 2014


Generalized linear models with functional predictors
journal, August 2002

  • James, Gareth M.
  • Journal of the Royal Statistical Society: Series B (Statistical Methodology), Vol. 64, Issue 3
  • DOI: 10.1111/1467-9868.00342

Functional single index models for longitudinal data
journal, February 2011

  • Jiang, Ci-Ren; Wang, Jane-Ling
  • The Annals of Statistics, Vol. 39, Issue 1
  • DOI: 10.1214/10-AOS845

Generalized Functional Linear Models With Semiparametric Single-Index Interactions
journal, June 2010

  • Li, Yehua; Wang, Naisyin; Carroll, Raymond J.
  • Journal of the American Statistical Association, Vol. 105, Issue 490
  • DOI: 10.1198/jasa.2010.tm09313

Functional Convex Averaging and Synchronization for Time-Warped Random Curves
journal, September 2004

  • Liu, Xueli; Müller, Hans-Georg
  • Journal of the American Statistical Association, Vol. 99, Issue 467
  • DOI: 10.1198/016214504000000999

Statistics of time warpings and phase variations
journal, January 2014

  • Marron, J. S.; Ramsay, James O.; Sangalli, Laura M.
  • Electronic Journal of Statistics, Vol. 8, Issue 2
  • DOI: 10.1214/14-EJS901

Functional Data Analysis of Amplitude and Phase Variation
journal, November 2015

  • Marron, J. S.; Ramsay, James O.; Sangalli, Laura M.
  • Statistical Science, Vol. 30, Issue 4
  • DOI: 10.1214/15-STS524

Functional Regression
journal, April 2015


On Estimating Regression
journal, January 1964

  • Nadaraya, E. A.
  • Theory of Probability & Its Applications, Vol. 9, Issue 1
  • DOI: 10.1137/1109020

Curve registration
journal, January 1998

  • Ramsay, J. O.; Li, Xiaochun
  • Journal of the Royal Statistical Society: Series B (Statistical Methodology), Vol. 60, Issue 2
  • DOI: 10.1111/1467-9868.00129

Methods for Scalar-on-Function Regression: Scalar-on-Function Regression
journal, February 2016

  • Reiss, Philip T.; Goldsmith, Jeff; Shang, Han Lin
  • International Statistical Review, Vol. 85, Issue 2
  • DOI: 10.1111/insr.12163

Consistent Estimation of Scaled Coefficients
journal, November 1986


Generative models for functional data using phase and amplitude separation
journal, May 2013


An exploratory NMR nutri-metabonomic investigation reveals dimethyl sulfone as a dietary biomarker for onion intake
journal, January 2009

  • Winning, Hanne; Roldán-Marín, Eduvigis; Dragsted, Lars O.
  • The Analyst, Vol. 134, Issue 11
  • DOI: 10.1039/b918259d

Dual tasking, gait rhythmicity, and Parkinson's disease: Which aspects of gait are attention demanding?
journal, September 2005


Estimator selection and combination in scalar-on-function regression
journal, February 2014


Wavelet-based scalar-on-function finite mixture regression models
journal, January 2016


An exploratory NMR nutri-metabonomic investigation reveals dimethyl sulfone as a dietary biomarker for onion intake
journal, January 2009

  • Winning, Hanne; Roldán-Marín, Eduvigis; Dragsted, Lars O.
  • The Analyst, Vol. 134, Issue 11
  • DOI: 10.1039/b918259d

Longitudinal scalar-on-functions regression with application to tractography data
journal, January 2013


Dual tasking, gait rhythmicity, and Parkinson's disease: Which aspects of gait are attention demanding?
journal, September 2005


Functional Data Analysis of Amplitude and Phase Variation
text, January 2015

  • Anuj, Srivastava,; O., Ramsay, James; M., Sangalli, Laura
  • The University of North Carolina at Chapel Hill University Libraries
  • DOI: 10.17615/yrhe-4052

Statistical Computing in Functional Data Analysis: The R Package fda.usc
journal, January 2012

  • Febrero-Bande, Manuel; Fuente, Manuel Oviedo de la
  • Journal of Statistical Software, Vol. 51, Issue 4
  • DOI: 10.18637/jss.v051.i04

Consistent Estimation of Scaled Coefficients
journal, November 1986


Generalized functional linear models
text, January 2005