MSinference-package {MSinference}R Documentation

Multiscale Inference for Nonparametric Time Trend(s)

Description

This package performs a multiscale analysis of a single nonparametric time trends (Khismatullina and Vogt (2020)) or multiple nonparametric time trends (Khismatullina and Vogt (2022), Khismatullina and Vogt (2023)).

In case of a single nonparametric regression, the multiscale method to test qualitative hypotheses about the nonparametric time trend m in the model Y_t = m(t/T) + \epsilon_t with time series errors \epsilon_t is provided. The method was first proposed in Khismatullina and Vogt (2020). It allows to test for shape properties (areas of monotonic decrease or increase) of the trend m.

This method require an estimator of the long-run error variance \sigma^2 = \sum_{l=-\infty}^{\infty} Cov(\epsilon_0, \epsilon_l). Hence, the package also provides the difference-based estimator for the case that the errors belong to the class of AR(\infty) processes. The estimator was also proposed in Khismatullina and Vogt (2020).

In case of multiple nonparametric regressions, we provide the multiscale method to test qualitative hypotheses about the nonparametric time trends in the context of epidemic modelling. Specifically, we assume that the we observe a sample of the count data \{\mathcal{X}_i = \{ X_{it}: 1 \le 1 \le T \}\}, where X_{it} are quasi-Poisson distributed with time-varying intensity parameter \lambda_i(t/T). The multiscale method allows to test whether intenisty parameters are different or not, and if they are, it detects with a prespicified significance level the regions where these differences most probably occur. The method was introduced in Khismatullina and Vogt (2023) and can be used for comparing the rates of infection of COVID-19 across countries.

References

Khismatullina M, Vogt M (2020). “Multiscale inference and long-run variance estimation in nonparametric regression with time series errors.” Journal of the Royal Statistical Society: Series B (Statistical Methodology).

Khismatullina M, Vogt M (2023). “Nonparametric comparison of epidemic time trends: The case of COVID-19.” Journal of Econometrics, 232(1), 87-108. ISSN 0304-4076, doi:10.1016/j.jeconom.2021.04.010.


[Package MSinference version 0.2.1 Index]