FFT-based fast bandwidth selector for multivariate kernel density estimation

被引：27

作者：

Gramacki, Artur ^{[1
]}

Gramacki, Jaroslaw ^{[2
]}

机构：

[1] Univ Zielona Gora, Inst Control & Computat Engn, Ul Licealna 9, PL-65417 Zielona Gora, Poland

[2] Univ Zielona Gora, Ctr Comp, Ul Licealna 9, PL-65417 Zielona Gora, Poland

来源：

COMPUTATIONAL STATISTICS & DATA ANALYSIS | 2017年 / 106卷

关键词：

Multivariate kernel density estimation; Density derivative functionals; Bandwidth selection; Fast Fourier Transform; Nonparametric estimation; CROSS-VALIDATION; CHOICE;

D O I：

10.1016/j.csda.2016.09.001

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

The performance of multivariate kernel density estimation (KDE) depends strongly on the choice of bandwidth matrix. The high computational cost required for its estimation provides a big motivation to develop fast and accurate methods. One of such methods is based on the Fast Fourier Transform. However, the currently available implementation works very well only for the univariate KDE and its multivariate extension suffers from a very serious limitation as it can accurately operate only with diagonal bandwidth matrices. A more general solution is presented where the above mentioned limitation is relaxed. Moreover, the presented solution can be easily adopted also for the task of efficient computation of integrated density derivative functionals involving an arbitrary derivative order. Consequently, bandwidth selection for kernel density derivative estimation is also supported. The practical usability of the new solution is demonstrated by comprehensive numerical simulations. (C) 2016 Elsevier B.V. All rights reserved.

引用

页码：27 / 45

页数：19

共 50 条

[41] Calibrated FFT-based density approximations for α-stable distributions
Menn, C
Rachev, VT
COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2006, 50 (08) : 1891 - 1904
[42] GRADIENT SCHEMES FOR ROBUST FFT-BASED MOTION ESTIMATION
Tzimiropoulos, Georgios
Argyriou, Vasileios
2016 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING PROCEEDINGS, 2016, : 1781 - 1785
[43] Fast multivariate empirical cumulative distribution function with connection to kernel density estimation
Langrene, Nicolas
Warin, Xavier
COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2021, 162
[44] BANDWIDTH SELECTION FOR KERNEL DENSITY-ESTIMATION
CHIU, ST
ANNALS OF STATISTICS, 1991, 19 (04): : 1883 - 1905
[45] Bootstrapping kernel spectral density estimates with kernel bandwidth estimation
Zoubir, AM
2003 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL VI, PROCEEDINGS: SIGNAL PROCESSING THEORY AND METHODS, 2003, : 325 - 328
[46] Bandwidth selection for kernel conditional density estimation
Bashtannyk, DM
Hyndman, RJ
COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2001, 36 (03) : 279 - 298
[47] A GPU acceleration for FFT-based fast solvers for the Integral Equation
Francavilla, M. A.
Attardo, E. A.
Vipiana, F.
Vecchi, G.
PROCEEDINGS OF THE FOURTH EUROPEAN CONFERENCE ON ANTENNAS AND PROPAGATION, 2010,
[48] Fast Kernel Distribution Function Estimation and fast kernel density estimation based on sparse Bayesian learning and regularization
Yin, Xun-Fu
Hao, Zhi-Feng
PROCEEDINGS OF 2008 INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND CYBERNETICS, VOLS 1-7, 2008, : 1756 - +
[49] Load Extrapolation Method Based on Adaptive Bandwidth Kernel Density Estimation
Niu W.
Cai F.
Fu J.
Nongye Jixie Xuebao/Transactions of the Chinese Society for Agricultural Machinery, 2021, 52 (01): : 375 - 384
[50] Progress in Data-Based Bandwidth Selection for Kernel Density Estimation
Comput Stat, 3 (337):

← 1 2 3 4 5 →