TY - JOUR
AU - Slaets, Leen
AU - Claeskens, Gerda
AU - Silverman, Bernard W.
PY - 2013/10/22
Y2 - 2022/12/04
TI - Warping Functional Data in R and C via a Bayesian Multiresolution Approach
JF - Journal of Statistical Software
JA - J. Stat. Soft.
VL - 55
IS - 3
SE - Articles
DO - 10.18637/jss.v055.i03
UR - https://www.jstatsoft.org/index.php/jss/article/view/v055i03
SP - 1 - 22
AB - Phase variation in functional data obscures the true amplitude variation when a typical cross-sectional analysis of these responses would be performed. Time warping or curve registration aims at eliminating the phase variation, typically by applying transformations, the warping functions τn, to the function arguments. We propose a warping method that jointly estimates a decomposition of the warping function in warping components, and amplitude components. For the estimation routine, adaptive MCMC calculations are performed and implemented in C rather than R to increase computational speed. The R-C interface makes the program user-friendly, in that no knowledge of C is required and all input and output will be handled through R. The R package MRwarping contains all needed files.
ER -