Title: R package for statistical process control when data are functional: fda.qcr
Authors: Miguel Flores - Escuela Politecnica Nacional (Ecuador)
Salvador Naya - University of A Coruna (Spain) [presenting]
Javier Tarrio-Saavedra - Universidade da Coruna (Spain)
Ruben Fernandez Casal - Universidade da Coruna (Spain)
Abstract: Functional Data Analysis, Quality Control and Reliability (fda.qcr) package is a new R library that tackles the problem of statistical process control from the Functional Data Analysis approach. It implements different FDA techniques to perform exploratory analysis, iterative visualization of functional data, outlier detection using data depth approach, analysis of variance, linear model fitting with scalar or binary response, and statistical quality control using specific FDA control charts. To obtain the confidence bands of control charts for functional data, the functional quantiles (i.e. 0.975 and 0.025) corresponding to critical to quality (CTQ) functional variable are obtained. Method based on data depth approach, pointwise functional quantiles estimates, and smoothed bootstrap resampling are implemented. FDA control charts can be applied taking into account dependence between observations of a functional CTQ variable. The fda.qcr package implements block bootstrap method to estimate functional quantiles when observations are dependent.