construct_weekly_grid

Computes the location-bandwidth weekly grid for the multiscale test.

Performs a multiscale analysis of a nonparametric
regression or nonparametric regressions with time series errors. In case
of one regression, with the help of this package it is possible to detect
the regions where the trend function is increasing or decreasing.
In case of multiple regressions, the test identifies regions where
the trend functions are different from each other. See
Khismatullina and Vogt (2020) <doi:10.1111/rssb.12347>,
Khismatullina and Vogt (2022) <doi:10.48550/arXiv.2209.10841> and
Khismatullina and Vogt (2023) <doi:10.1016/j.jeconom.2021.04.010>
for more details on theory and applications.

Marina Khismatullina

MSinference

Multiscale Inference for Nonparametric Time Trend(s)

Michael Vogt

construct_weekly_grid function

<dl><dt>t</dt>
<dd>Sample size.</dd>
<dt>min_len</dt>
<dd>Minimal length of the interval considered. The grid then
consists of intervals with lengths min_len,
2 * min_len, 3 * min_len, ... Default is 7, i.e. a week.</dd>
<dt>nmbr_of_wks</dt>
<dd>Number that determines the longest intervals in the grid:
the length of this interval is calculated then as 
min_len * nmbr_of_wks. Default is 4.</dd></dl>

Arguments

Computes the location-bandwidth weekly grid for the multiscale test. — construct_weekly_grid

<dl>

<dt>t</dt>
<dd>Sample size.</dd>


<dt>min_len</dt>
<dd>Minimal length of the interval considered. The grid then
consists of intervals with lengths min_len,
2 * min_len, 3 * min_len, ... Default is 7, i.e. a week.</dd>


<dt>nmbr_of_wks</dt>
<dd>Number that determines the longest intervals in the grid:
the length of this interval is calculated then as 
min_len * nmbr_of_wks. Default is 4.</dd>

</dl>

construct_weekly_grid: Computes the location-bandwidth weekly grid for the multiscale test.

Description

Usage

Value

Arguments

Examples