Qtools (version 1.5.8)

midq2q: Recover Ordinary Conditional Quantiles from Conditional Mid-Quantiles

Description

This function recovers ordinary conditional quantile functions based on fitted mid-quantile regression models.

Usage

midq2q(object, newdata, observed = FALSE, ...)

Value

a vector or a matrix of estimated ordinary quantiles. The attribute Fhat provides the corresponding estimated cumulative distribution.

Arguments

object

an object of class midrq.

newdata

a required data frame in which to look for variables with which to predict.

observed

logical flag. If TRUE, ordinary quantiles are recovered from observed sample values. Otherwise, they are calcuated as rounded mid-quantiles. See details.

...

not used.

Author

Marco Geraci

Details

If the values of the support of the random variable are equally spaced integers, then observed should ideally be set to FALSE so that the ordinary quantile is obtained by rounding the predicted mid-quantile. Otherwise, the function returns an integer observed in the sample. See Geraci and Farcomeni for more details.

References

Geraci, M. and A. Farcomeni. Mid-quantile regression for discrete responses. arXiv:1907.01945 [stat.ME]. URL: https://arxiv.org/abs/1907.01945.

See Also

plot.midq2q, predict.midrq

Examples

if (FALSE) {
# Esterase data
data(esterase)

# Fit quantiles 0.1, 0.15, ..., 0.85
fit <- midrq(Count ~ Esterase, tau = 2:17/20, data = esterase, type = 3, lambda = 0)

# Recover ordinary quantile function
xx <- seq(min(esterase$Esterase), max(esterase$Esterase), length = 5)
print(Qhat <- midq2q(fit, newdata = data.frame(Esterase = xx)))

# Plot
plot(Qhat, sub = TRUE)

}