The opportunity to utilize multivariate functional data types for conducting classification tasks is emerging with the growing availability of imaging data. Inspired by the extensive data provided by the Alzheimer’s Disease Neuroimaging Initiative, we introduce a novel classifier tailored for multivariate functional data. Each observation in this framework may be associated with numerous functional processes, varying in dimensions, such as curves and images. Each predictor is a random element in an infinite-dimensional function space, and the number of functional predictors can potentially be much greater than the sample size. By adopting a sparse deep rectified linear unit network architecture and incorporating the LassoNet algorithm, the proposed functional Bayesian information criterion deep neural network performs feature selection and classification simultaneously, in contrast to existing functional data classifiers. This approach addresses the challenge of complex inter-correlation structures among multiple functional processes without requiring distributional assumptions. A simulation study and a real data application demonstrate its favorable performance.
LiCXiaoLLuoS. Joint model for survival and multivariate sparse functional data with application to a study of Alzheimer’s disease. Biometrics2022; 78: 435–447.
2.
WangXNanBZhuJ, et al.Regularized 3D functional regression for brain image data via haar wavelets. Ann Appl Stat2014; 8: 1045.
3.
WangXNanBZhuJ, et al.Classification of ADNI pet images via regularized 3D functional data analysis. Biostat Epidemiol2017; 1: 3–19.
4.
MüllerH-g. Functional modelling and classification of longitudinal data. Scand J Stat2005; 32: 223–240.
5.
JamesGMHastieT. Functional linear discriminant analysis for irregularly sampled curves. J R Stat Soc B2001; 63: 533–550.
6.
DelaigleAHallP. Achieving near-perfect classification for functional data. J R Stat Soc B2012; 74: 267–286.
7.
CérouFGuyaderA. Nearest neighbor classification in infinite dimension. ESAIM2006; 10: 340–355.
8.
DaiWGentonMG. An outlyingness matrix for multivariate functional data classification. Stat Sin2018; 28: 2435–2454.
9.
MoindjiéI-ADabo-NiangSPredaC. Classification of multivariate functional data on different domains with partial least squares approaches. Stat Comput2024; 34: 5.
10.
YaoJMuellerJWangJ-L. Deep learning for functional data analysis with adaptive basis layers. In: International conference on machine learning, 2021, pp. 11898–11908. PMLR.
ThindBMultaniKCaoJ. Deep learning with functional inputs. J Comput Graph Stat2023; 32: 171–180.
13.
WuSBeaulacCCaoJ. Neural networks for scalar input and functional output. Stat Comput2023; 33: 118.
14.
WangSCaoGShangZ. Deep neural network classifier for multi-dimensional functional data. Scand J Stat2023; 50: 1667–1686.
15.
WangSShangZCaoG, et al.Optimal classification for functional data. Stat Sin2024; 34: 1545–1564.
16.
WangSHuangYCaoG. Review on functional data classification. WIREs Comput Stat2023; 16: e1638.
17.
BerrenderoJRCuevasATorrecillaJL. Variable selection in functional data classification: a maxima-hunting proposal. Stat Sin2016; 7: 619–638.
18.
YuWWadeSBondellDH, et al.Nonstationary Gaussian process discriminant analysis with variable selection for high-dimensional functional data. J Comput Graph Stat2023; 32: 588–600.
19.
HeKZhangXRenS, et al.Deep residual learning for image recognition. In: Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–778.
20.
LinHJegelkaS. Resnet with one-neuron hidden layers is a universal approximator. In: Advances in neural information processing systems, 2018, pp. 6169–6178.
21.
Schmidt-HieberJ. Nonparametric regression using deep neural networks with ReLU activation function. Ann Stat2020; 48: 1875–1897.
22.
WangSCaoG. Multiclass classification for multidimensional functional data through deep neural networks. Electron J Stat2024; 18: 1248–1292.
23.
LemhadriIRuanFAbrahamL, et al.Lassonet: a neural network with feature sparsity. J Mach Learn Res2021; 130: 10–18.
24.
WangLKimYLiR. Calibrating non-convex penalized regression in ultra-high dimension. Ann Stat2013; 41: 2505–2536.
25.
YaoFMüllerHGWangJL. Functional data analysis for sparse longitudinal data. J Am Stat Assoc2005; 100: 577–590.
26.
LiJFineJP. Roc analysis with multiple tests and multiple classes: methodology and applications in microarray studies. Biostatistics2008; 9: 566–576.
27.
LiJGaoMD’AgostinoRB. Evaluating classification accuracy for modern learning approaches. Stat Med2019; 38: 2477–2503. Tutorials in Biostatistics.
28.
LeCunYBengioYHintonG. Deep learning. Nature2015; 521: 436–444.
29.
AudibertJ-YTsybakovAB. Fast learning rates for plug-in classifiers. Ann Stat2007; 35: 608–633.
30.
BosTSchmidt-HieberJ. Convergence rates of deep ReLU networks for multiclass classification. Electron J Stat2022; 16: 2724–2773.
31.
Alzheimer’s Disease Neuroimaging Initiative (ADNI). Cognitive assessments and questionnaires (ADNI quick-start guide). Online (accessed 16 February 2026).
32.
CedarbaumJMJarosMHernandezC, et al.Rationale for use of the clinical dementia rating sum of boxes as a primary outcome measure for Alzheimer’s disease clinical trials. Alzheimers Dement2013; 9: S45–S55.
33.
ADNI Reference Documentation b. ADAS-cognitive (ADAS)—variable definitions used in ADNI-derived datasets. Online (accessed 16 February 2026).
34.
Alzheimer’s Disease Neuroimaging Initiative (ADNI). aDNI3 procedures manual (version 3.0). Online PDF (2017, accessed 16 February 2026).
35.
Alzheimer’s Disease Neuroimaging Initiative (ADNI). ADNI go procedures manual. Online PDF (2011, accessed 16 February 2026).
36.
ADNI Reference Documentation a. The preclinical Alzheimer’s cognitive composite (PACC) and modified PACC scores. Online (accessed 16 February 2026).
