Measurement errors of medico-technical devices can be separated into systematic bias and random error. We propose a new method to address both simultaneously via generalized additive models for location, scale and shape (GAMLSS) in combination with permutation tests. More precisely, we extend a recently proposed boosting algorithm for GAMLSS to provide a test procedure to analyse potential device effects on the measurements. We carried out a large-scale simulation study to provide empirical evidence that our method is able to identify possible sources of systematic bias as well as random error under different conditions. Finally, we apply our approach to compare measurements of skin pigmentation from two different devices in an epidemiological study.
AtkinsonGNevillAM. Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Med1998; 26(4): 217–238.
2.
UterWBenzMMayrA. Assessing skin pigmentation in epidemiological studies: The reliability of measurements under different conditions. Skin Res Technol2013; 19(2): 100–106.
3.
RigbyRAStasinopoulosD. Generalized additive models for location, scale and shape (with discussion). Appl Statist2005; 54: 507–554.
RigbyRAStasinopoulosDM. Automatic smoothing parameter selection in GAMLSS with an application to centile estimation. Statist Meth Med Res2014; 23(4): 346–368.
6.
Klein N, Kneib T and Lang S. Bayesian generalised additive models for location, scale, and shape for zero-inflated and overdispersed count data. Journal of the American Statistical Association (JASA) 2015. DOI:10.1080/01621459.2014.912955.
7.
MayrAFenskeNHofnerB. Generalized additive models for location, scale and shape for high-dimensional data – a flexible approach based on boosting. J R Statist Soc Ser C Appl Statist2012; 61(3): 403–427.
8.
KneibT. Beyond mean regression. Statist Modell2013; 13(4): 275–303.
9.
FenskeNKneibTHothornT. Identifying risk factors for severe childhood malnutrition by boosting additive quantile regression. JASA2011; 106(494): 494–510.
10.
HothornTKneibTBühlmannP. Conditional transformation models. J R Statist Soc Ser B Statist Methodol2014; 76(1): 3–27.
11.
van BuurenS. Growth charts of human development. Statist Meth Med Res2014; 23(4): 346–368.
12.
van DommelenPvan BuurenS. Methods to obtain referral criteria in growth monitoring. Statist Meth Med Res2014; 23(4): 369–389.
13.
De Onis M and WHO Multicentre Growth Reference Study Group. WHO child growth standards: length/height-for-age, weight-for-age, weight-for-length, weight-for-height and body mass index-for-age: Methods and development. Geneva: World Health Organization 2006; Available at: http://www.who.int/iris/handle/10665/43413.
14.
WahlSFenskeNZeilingerS. On the potential of models for location and scale for genome-wide DNA methylation data. BMC Bioinform2014; 15(1): 232–232.
15.
Muniz-Terrera G, van den Hout A, Rigby R et al. Analysing cognitive test data: Distributions and non-parametric random effects. Statist Meth Med Res 2016; 25:741–753.
16.
BühlmannPHothornT. Boosting algorithms: Regularization, prediction and model fitting (with discussion). Statist Sci2007; 22: 477–522.
17.
HofnerBMayrARobinzonovN. Model-based boosting in R: A hands-on tutorial using the R package mboost. Computat Statist2014; 29(1–2): 3–35.
SchmidMPotapovSPfahlbergA. Estimation and regularization techniques for regression models with multidimensional prediction functions. Statist Comput2010; 20: 139–150.
20.
StasinopoulosDRigbyR. Generalized additive models for location scale and shape (GAMLSS) in R. J Statist Softw2007; 23(7): 1–46.
21.
MayrABinderHGefellerO. The evolution of boosting algorithms - from machine learning to statistical modelling. Meth Inform Med2014; 53(6): 419–427.
22.
MayrABinderHGefellerO. Extending statistical boosting - an overview of recent methodological developments. Meth Inform Med2014; 53(6): 428–435.
23.
HastieTTibshiraniRFriedmanJ. The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd edn, New York: Springer, 2009.
24.
MayrAHofnerBSchmidM. The importance of knowing when to stop – a sequential stopping rule for component-wise gradient boosting. Meth Inform Med2012; 51(2): 178–186.
25.
R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, 2014. http://www.R-project.org/.
26.
Hofner B, Mayr A and Schmid M. gamboostlss: An R package for model building and variable selection in the gamlss framework. Preprint arXiv:14071774, 2014.
27.
DraperNRStonemanDM. Testing for the inclusion of variables in linear regression by a randomisation technique. Technometrics1966; 8(4): 695–699.
PotterDM. A permutation test for inference in logistic regression with small and moderate-sized data sets. Statist Med2005; 24(5): 693–708.
30.
WerftWBennerA. glperm: A permutation of regressor residuals test for inference in generalized linear models. The R Journal2010; 2(1): 39–43.
31.
BoulesteixALHothornT. Testing the additional predictive value of high-dimensional molecular data. BMC Bioinformatics2010; 11: 78–78. DOI:10.1186/1471-2105-11-78.
32.
Borges P, Rodrigues J, Louzada F et al. A cure rate survival model under a hybrid latent activation scheme. Statist Meth Med Res 2016; 25: 838–856.
33.
FitzpatrickTB. The validity and practicality of Sun-reactive skin types I through VI. Arch Dermatol1988; 124(6): 869–871.
34.
FiesslerCPfahlbergALiJ. Accuracy and reliability of naevus self-counts. Melanoma Res2014; 24(6): 611–616.
35.
BühlmannPGertheissJHiekeS. Discussion of “The evolution of boosting algorithms” and “Extending statistical boosting”. Meth Inf Med2014; 53: 436–445.
