broom (version 0.4.1)

rq_tidiers: Tidying methods for quantile regression models

Description

These methods tidy the coefficients of a quantile regression model into a summary, augment the original data with information on the fitted values and residuals, and construct a glance of the model's statistics.

Usage

"tidy"(x, se.type = "rank", conf.int = TRUE, conf.level = 0.95, alpha = 1 - conf.level, ...)
"tidy"(x, se.type = "rank", conf.int = TRUE, conf.level = 0.95, alpha = 1 - conf.level, ...)
"tidy"(x, conf.int = FALSE, conf.level = 0.95, ...)
"glance"(x, ...)
"glance"(x, ...)
"augment"(x, data = model.frame(x), newdata, ...)
"augment"(x, data = model.frame(x), newdata, ...)
"augment"(x, data = NULL, newdata = NULL, ...)

Arguments

x

model object returned by rq or nlrq

se.type

Type of standard errors to calculate; see summary.rq

conf.int

boolean; should confidence intervals be calculated, ignored if se.type = "rank"

conf.level

confidence level for intervals

alpha

confidence level when se.type = "rank"; defaults to the same as conf.level although the specification is inverted

data

Original data, defaults to extracting it from the model

newdata

If provided, new data frame to use for predictions

...

other arguments passed on to summary.rq

Value

All tidying methods return a data.frame without rownames, whose structure depends on the method chosen.tidy.rq returns a data frame with one row for each coefficient. The columns depend upon the confidence interval method selected.tidy.rqs returns a data frame with one row for each coefficient at each quantile that was estimated. The columns depend upon the confidence interval method selected.tidy.nlrq returns one row for each coefficient in the model, with five columns:

term

The term in the nonlinear model being estimated and tested

estimate

The estimated coefficient

std.error

The standard error from the linear model

statistic

t-statistic

p.value

two-sided p-value

glance.rq returns one row for each quantile (tau) with the columns:

tau

quantile estimated

logLik

the data's log-likelihood under the model

AIC

the Akaike Information Criterion

BIC

the Bayesian Information Criterion

df.residual

residual degrees of freedom

glance.rq returns one row for each quantile (tau) with the columns:

tau

quantile estimated

logLik

the data's log-likelihood under the model

AIC

the Akaike Information Criterion

BIC

the Bayesian Information Criterion

df.residual

residual degrees of freedom

augment.rq returns a row for each original observation with the following columns added:

.resid

Residuals

.fitted

Fitted quantiles of the model

.tau

Quantile estimated

Depending on the arguments passed on to predict.rq via ... a confidence interval is also calculated on the fitted values resulting in columns:

.conf.low

Lower confidence interval value

.conf.high

Upper confidence interval value

See predict.rq for details on additional arguments to specify confidence intervals. predict.rq does not provide confidence intervals when newdata is provided.augment.rqs returns a row for each original observation and each estimated quantile (tau) with the following columns added:

.resid

Residuals

.fitted

Fitted quantiles of the model

.tau

Quantile estimated

predict.rqs does not return confidence interval estimates.augment.rqs returns a row for each original observation with the following columns added:

.resid

Residuals

.fitted

Fitted quantiles of the model

Details

If se.type != "rank" and conf.int = TRUE confidence intervals are calculated by summary.rq. Otherwise they are standard t based intervals.

This simply calls augment.nls on the "nlrq" object.