# 72_quantile-based_statistics: Robust Statistics In probhat: Multivariate Generalized Kernel Smoothing and Related Statistical Methods

## Description

Compute robust (quantile based) statistics from probability distributions.

## Usage

 ```1 2 3 4``` ```ph.median (xf, ...) ph.quantile (xf, p, ...) iqr (xf, P=0.5, ...) ```

## Arguments

 `xf` A numeric vector, suitable function object, or an object that can be coerced to a numeric vector. Here, a suitable function object is a quantile function. Refer to the references and see also sections. `P, p` Numeric vectors, the probabilities. P is the area (probability) between the lower and upper limits. `...` Other arguments. Refer to the details section.

## Details

If xf is a numeric vector, a qfuv.el object is created using xf as the main argument.
Any arguments contained within ..., are passed to the qfuv.el constructor.

If xf is not a quantile function, these functions try to coerce it to a numeric vector, and apply the above.

## Value

ph.median returns a single numeric value.

The other functions return a numeric vector.

## References

Refer to the vignette for an overview, references and better examples.

Succinct Constructors
Discrete Kernel Smoothing, Continuous Kernel Smoothing, Empirical-Like Distributions

probmv, rng

ph.mean, moment
quartiles, ntiles
ph.mode, ph.modes

## Examples

 ```1 2 3 4 5 6 7``` ```prep.ph.data () cFht <- qfuv.cks (height) cFht (0.5) ph.median (cFht) #iqr (cFht) ```

probhat documentation built on May 12, 2021, 5:08 p.m.