plotQGAMs: Plot quantiles

In PeterHendrix13/distWorkshop: Distributional analyses of linguistic data in the GAMM framework

Description

Plot the results of a series of QGAM models (Fasiolo et al., 2017)

Usage

plotQGAMs(models, predictor, data, cols = c("#000080", "#1A1A9A",
  "#3333B3", "#4D4DCD", "#6666E6"), se = 2, xlab = NA, ylab = NA,
  main = NA, ylim = NA, ...)

Arguments

models	A list of QGAM models as generated by the mqgam() function in the qgam package.
predictor	The predictor to be plotted. This predictor needs to be present in the fitted models, as well as in data.
data	The data the QGAM models were fit to. Needs to include the response variable, as well as all predictors in these models.
cols	A vector of colors. The lines corresponding to the quantiles will be plotted in these colors.
se	The number of standard errors for the confidence intervals. Default: 2 (i.e., 95% confidence intervals)

References

Fasiolo M., Goude Y., Nedellec R., & Wood S. N. (2017). Fast calibrated additive quantile regression. URL: https://arxiv.org/abs/1707.03307.

Keuleers, E., Lacey, P., Rastle, K., & Brysbaert, M. (2012). The British Lexicon Project: Lexical decision data for 28,730 monosyllabic and disyllabic English words. Behavior Research Methods, 44(1), 287-304.

Examples

# Remove outliers from the ld data set, which contains lexical 
# decision latencies from the British Lexicon Project (Keuleers 
# et al, 2012)
predictors = c("RT", "logFrequency", "Length", "logOLD20", "SND20")
ld = removeOutliers(ld, predictors)
ld = na.omit(ld)

# Tune learning rate for median
tune = tuneLearnFast(RT ~ s(logFrequency) + s(Length) + 
                          s(logOLD20) + s(SND20),
                     data = ld, qu = 0.5)
sigpar = tune$lsig

# Define quantiles
quants = c(0.10,0.25,0.50,0.75,0.90)

# Run qgam models
qgams = mqgam(RT ~ s(logFrequency) + s(Length) + s(logOLD20) + 
                   s(SND20),
                   data = ld, qu = quants, lsig = sigpar)

# Plot effect of frequency at quantiles
plotQGAMs(qgams, "logFrequency", ld)

PeterHendrix13/distWorkshop documentation built on Nov. 5, 2019, 2:51 p.m.