rqcomb: Combined quantile regression

In gbstat/tailqr: 'Fit Mixture of Quantile Regression and GPDs'

Description

Fit different quantile regression models to the bulk quantiles and the tail quantiles. This allows for a higher order basis design in the bulk quantiles and a lower order basis design in the tails.

Usage

rqcomb(y, dates, b.df.lf, b.df.hf, b.b.formula, t.df.lf, t.df.hf, t.b.formula,
  taus, l.taus, u.taus, pred.dates = NULL)

Arguments

y	response vector
dates	vector of dates of each observation
b.df.lf	df for bulk low freq basis
b.df.hf	df for bulk high freq basis
b.b.formula	bulk formula involving b.lf and b.hf
t.df.lf	df for tail low freq basis
t.df.hf	df for tail high freq basis
t.b.formula	tail formula (simpler) involving b.lf and b.hf
taus	vector of intermediate (bulk) quantiles to estimate (between 0 and 1)
l.taus	vector of lower tail quantiles to estimate (between 0 and 1)
u.taus	vector of higher tail quantiles to estimate (between 0 and 1)
pred.dates	(optional) vector of dates to make predictions on (if different from dates)

Value

List of lower, mid, and upper quantile estimates

gbstat/tailqr documentation built on May 8, 2019, 5:42 p.m.