wqrr — Wavelet Quantile Regression Toolbox

Author / maintainer: Dr Merwan Roudane  ·  merwanroudane920@gmail.com  ·  Repo: https://github.com/merwanroudane/wqrr

A comprehensive toolbox for wavelet-domain quantile analyses of bivariate and multivariate time series. Eight estimators in one consistent interface, MODWT decomposition via waveslim, and interactive plotly visualisations that default to MATLAB Parula.

Function Method
wavelet_qr() Wavelet Quantile Regression — band x quantile slopes
multivariate_wqr() Multivariate WQR — multiple regressors per band
wavelet_qqr() Wavelet QQR — (theta, tau) coefficient + p-value surfaces
np_quantile_causality() Nonparametric Causality-in-Quantiles
wavelet_np_causality() Wavelet variant of the above
wavelet_mediation() Direct, interaction, paths a / b, indirect
wavelet_quantile_correlation() Wavelet Quantile Correlation with CIs
wavelet_quantile_density() Wavelet nonparametric quantile density

For plain bivariate Quantile-on-Quantile regression see the companion CRAN package QuantileOnQuantile.

Installation

# from CRAN (once accepted)
install.packages("wqrr")

# development version
# install.packages("remotes")
# remotes::install_github("merwanroudane/wqrr")
install.packages("wqrr", repos = c('https://merwanroudane.r-universe.dev', 'https://cloud.r-project.org'))

# from a local source tarball
install.packages("wqrr_1.0.0.tar.gz", repos = NULL, type = "source")

60-second tour

library(wqrr)

dat <- read.csv(system.file("extdata", "wqrr_data.csv", package = "wqrr"))

# Wavelet QR: S&P 500 returns on oil returns
wqr_fit <- wavelet_qr(dat$sp500_return, dat$oil_return,
                      quantiles = seq(0.1, 0.9, by = 0.1),
                      wavelet = "la8", J = 5)
print(wqr_fit)
plot(wqr_fit)                # MATLAB Parula by default

# Wavelet QQR surface (theta x tau) with p-values
wqqr_fit <- wavelet_qqr(dat$sp500_return, dat$oil_return,
                        wavelet = "la8", J = 5, band = "long")
plot(wqqr_fit, type = "3d")

# Causality-in-quantiles
cause_fit <- np_quantile_causality(dat$oil_return, dat$sp500_return)
plot(cause_fit)

# Mediation / moderation via EPU
med <- wavelet_mediation(dat$sp500_return, dat$oil_return, dat$epu,
                         wavelet = "la8", J = 5)
panels <- plot_mediation_panel(med)
panels$Indirect

See vignette("wqrr-introduction") for the full walkthrough.

Colour palettes

wqrr_colorscales()
#> Parula           : MATLAB R2014b default (perceptually uniform)
#> Jet              : Classic MATLAB rainbow
#> Turbo            : Google Turbo
#> BlueRed          : Diverging (blue low, red high)
#> Sinha            : Sinha cross-quantile heatmap
#> GreenOrangeRed   : WQR-style sequential
#> GreenYellowRed   : Mediation-panel default
#> Viridis, Plasma, Cividis, Inferno, Magma, RdBu (plotly built-ins)

References

License

GPL-3