Special Issue on HIGH-DIMENSIONAL AND FUNCTIONAL DATA ANALYSIS - Archive ouverte HAL Accéder directement au contenu
N°Spécial De Revue/Special Issue Computational Statistics and Data Analysis Année : 2019

Special Issue on HIGH-DIMENSIONAL AND FUNCTIONAL DATA ANALYSIS

Résumé

High-dimensional and functional data structures have been an important focus of methodological, theoretical and applied statistics research for over two decades, with applications areas including biostatistics, chemometrics, environmetrics, genetics, geophysics, and neuroimaging. The aim of this special issue is to collect research papers concerned with computational and data-analytic aspects of high-dimensional and functional data analysis. The papers collected in this issue encompass a broad range of topics and show how much these areas have expanded. They show that many challenging problems remain, ranging from fundamental methodological questions to specialized applications. Many papers open up new directions of research. Applications studied in the papers include biology, climate and environmental research, genomics, finance, medical research, electric power market and road traffic. Due to the scope of topics covered in this issue, any attempt to organize them into natural groups must be somewhat arbitrary, but we will attempt to do so to offer the reader an overview of the contents. We focus on statistical methodology rather than the specific areas of applications listed in the previous paragraph. Classification and clustering have occupied a central role in both areas of statistics represented in this issue. The papers of Abpeykar, Ghatee and Zare (2018) and Park and Simpson (2018) make contributions, respectively, to high-dimensional and functional classification. Related to classification are the concepts of depth and outlier detection, which are particularly hard in the context of functional data, as there are at the same time many ways to order the data and none of them is obviously natural. The papers of Dai and Genton (2018) and Martinez-Hernandez, Genton and Gonzalez-Farias (2018) study multivariate and dependent functional data from these angles. An established direction of functional data analysis has been the study of various aspect of the many types of regression models. This issue contains three papers that provide in-depth insights and new methodology related to functional regression. Wang and Xu (2018) develop new methodology for nonparametric functional regression. Ma, Li, Zhu and Zhu (2018) extend quantile regression to high-dimensional functional regression. Febrero-Bande, Galeano and Gonzlez-Manteiga (2018) study regression with missing responses. While the paper of Liebl and Rameseder (2018) does not focus on regression, it is also concerned with a special form of missing data. The paper of Ahmed (2018) considers a modified RV coefficient, which aims at measuring linear dependence between two mul-tivariate data sets, that better suits high-dimensional data and proposes a new test of significance for the true coefficient. Penalized estimation and registration are important directions of functional data analysis, which are represented by the papers of Wong and Zhang (2018), Zhang, Zou, Ravishanker and Thavaneswaran (2018) and Fu and Heckman (2018). Several papers do not fall into "traditional" directions. The paper of French, Kokoszka, Stoev and Hall (2018) combines the methods of functional data analysis and extreme value theory to study the risk of heat waves. Sang, Wang and Cao (2018) apply empirical likelihood to dynamical correlations. Shen, Yao and Li (2018) study the estimation of a large volatility matrix from high-frequency data. The paper of Yue, Li and Cheng (2018) considers boosting in the context of high-dimensional longitudinal data. We hope that the papers published in this issue show the breadth, depth and vitality of the areas of high-dimensional and functional data analysis, and will stimulate continued research in these fields.
Fichier non déposé

Dates et versions

hal-01981010 , version 1 (28-01-2019)

Identifiants

  • HAL Id : hal-01981010 , version 1

Citer

Frédéric Ferraty, Piotr Kokoszka, Jane-Ling Wang, Yichao Wu. Special Issue on HIGH-DIMENSIONAL AND FUNCTIONAL DATA ANALYSIS. Computational Statistics and Data Analysis, 2019, Special Issue on High-dimensional and Functional Data Analysis. ⟨hal-01981010⟩
235 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More