Abstract:We consider strictly stationary stochastic processes of Hilbert space-valued random variables and focus on fully functional tests for the equality of the lag-zero autocovariance operators of several independent functional time series. A moving block bootstrap-based testing procedure is proposed which generates pseudo random elements that satisfy the null hypothesis of interest. It is based on directly bootstrapping the time series of tensor products which overcomes some common difficulties associated with applications of the bootstrap to related testing problems. The suggested methodology can be potentially applied to a broad range of test statistics of the hypotheses of interest. As an example, we establish validity for approximating the distribution under the null of a test statistic based on the Hilbert-Schmidt distance of the corresponding sample lag-zero autocovariance operators, and show consistency under the alternative. As a prerequisite, we prove a central limit theorem for the moving block bootstrap procedure applied to the sample autocovariance operator which is of interest on its own. The finite sample size and power performance of the suggested moving block bootstrap-based testing procedure is illustrated through simulations and an application to a real-life dataset is discussed.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to test whether the lag - zero auto - covariance operators are equal in multiple independent functional time series. Specifically, researchers are concerned with how to evaluate the differences between different sample lag - zero auto - covariance operators through fully functional test statistics, and these differences are measured by appropriate distance metrics. ### Background and Problem Description of the Paper In functional data analysis, random variables can be curves or images and can be represented in an appropriate function space. This paper considers functional time series \(X_n=\{X_1, X_2,\ldots, X_n\}\) generated from a strictly stationary random process \(X = \{X_t, t\in\mathbb{Z}\}\), where \(X_t(\tau)\) is a random function in a Hilbert space. It is assumed that these functions are \(L^4\)-m - approximable, which is a dependence assumption applicable to many commonly used functional time - series models. Researchers hope to infer the properties of a set of \(K\) independent functional processes based on the observed data of each group. In particular, the problem they are concerned with is to test whether the lag - zero auto - covariance operators of these \(K\) processes are equal, and consider using fully functional test statistics to evaluate the differences between the corresponding sample lag - zero auto - covariance operators. ### Solution To implement the proposed test method, researchers applied a Moving Block Bootstrap (MBB) procedure to estimate the distribution of the test statistic under the null hypothesis. The specific steps are as follows: 1. **Calculate the pooled mean**: \[ Y_M=\frac{1}{M}\sum_{i = 1}^K\sum_{t = 1}^{n_i}\hat{Y}_{i,t} \] where \(\hat{Y}_{i,t}=(X_{i,t}-\bar{X}_{i,n_i})\otimes(X_{i,t}-\bar{X}_{i,n_i})\), and \(\bar{X}_{i,n_i}\) is the sample mean function of the \(i\)-th group. 2. **Calculate the block mean**: \[ \tilde{Y}_{i,\xi}=\frac{1}{N_i}\sum_{t=\xi}^{N_i+\xi - 1}\hat{Y}_{i,t},\quad\xi = 1,2,\ldots,b_i \] 3. **Generate bootstrap pseudo - observations**: \[ Y^*_{i,t}=Y_M+\hat{Y}^*_{i,t}-\tilde{Y}_{i,\xi} \] where \(\hat{Y}^*_{i,\xi+(s - 1)b_i}=\hat{Y}_{i,q_{i,s}+\xi - 1}\), and \(q_{i,s}\) is an integer independently drawn from a discrete uniform distribution. 4. **Calculate the bootstrap statistic**: \[ T^*_M=\frac{n_1n_2}{M}\|Y^*_{1,n_1}-Y^*_{2,n_2}\|^2_{HS} \] where \(Y^*_{i,n_i}=\frac{1}{n_i}\sum_{t = 1}^{n_i}Y^*_{i,t}\). ### Results and Discussion Through simulation experiments, researchers verified the effectiveness of this method under different dependence structures. The results show that the MBB - based test procedure can maintain the nominal level, and as the degree of deviation from the null hypothesis increases, the power of the test also increases. In addition, researchers also applied this method to the actual daily temperature data in Nicosia, Cyprus in the summers of 2007 and 2009, demonstrating the applicability of this method in real - data.

Testing Equality of Autocovariance Operators for Functional Time Series

Specification Testing in Nonstationary Time Series Models

Testing Equality of Spectral Density Operators for Functional Processes

Testing for Independence Between Two Covariance Stationary Time Series

Testing for Independence Between Two Stationary Time Series Via the Empirical Characteristic Function

A General Approach to Testing Volatility Models in Time Series

Specification Tests for Time-Varying Coefficient Models

Testing for Pairwise Serial Independence Via the Empirical Distribution Function

Testing equality between several populations covariance operators

Diagnostic Checking for the Adequacy of Nonlinear Time Series Models

A Blockwise Bootstrap-Based Two-Sample Test for High-Dimensional Time Series

Testing marginal homogeneity in Hilbert spaces with applications to stock market returns

A Global Wavelet Based Bootstrapped Test of Covariance Stationarity

Testing Separability of Functional Time Series

Hypothesis Testing in Time Series Via the Empirical Characteristic Function: A Generalized Spectral Density Approach

Approach to Testing Nonlinear Time Series Models

An Introduction to Bootstrap Theory in Time Series Econometrics

A bootstrap test for equality of variances

Bootstrap Inference in Cointegrating Regressions: Traditional and Self-Normalized Test Statistics

Generalized Spectral Tests for Conditional Mean Models in Time Series with Conditional Heteroscedasticity of Unknown Form

Inference for the cross-covariance operator of stationary functional time series