# binning: Linear binning In npsp: Nonparametric Spatial Statistics

## Description

Discretizes the data into a regular grid (computes a binned approximation) using the multivariate linear binning technique described in Wand (1994).

## Usage

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14``` ```binning(x, y = NULL, nbin = NULL, set.NA = FALSE) as.bin.data(object, ...) ## S3 method for class 'data.grid' as.bin.data(object, data.ind = 1, weights.ind = NULL, ...) ## S3 method for class 'bin.data' as.bin.data(object, ...) ## S3 method for class 'SpatialGridDataFrame' as.bin.data(object, data.ind = 1, weights.ind = NULL, ...) ```

## Arguments

 `x` vector or matrix of covariates (e.g. spatial coordinates). Columns correspond with covariates (coordinate dimension) and rows with data. `y` vector of data (response variable). `nbin` vector with the number of bins on each dimension. `set.NA` logical. If `TRUE`, sets the bin averages corresponding to cells without data to `NA`. `object` (gridded data) used to select a method. `...` further arguments passed to or from other methods. `data.ind` integer (or character) with the index (or name) of the component containing the bin averages. `weights.ind` integer (or character) with the index (or name) of the component containing the bin counts/weights (if not specified, they are set to `as.numeric( is.finite( object[[data.ind]] ))`).

## Details

If parameter `nbin` is not specified is set to `pmax(25, rule.binning(x))`.

Setting `set.NA = TRUE` (equivalent to `biny[binw == 0] <- NA`) may be useful for plotting the binned averages `\$biny` (the hat matrix should be handled with care when using `locpol`).

## Value

If `y != NULL`, an S3 object of `class` `bin.data` (gridded binned data; extends `bin.den`) is returned. A `data.grid` object with the following 4 components:

 `biny` vector or array (dimension `nbin`) with the bin averages. `binw` vector or array (dimension `nbin`) with the bin counts (weights). `grid` a `grid.par`-`class` object with the grid parameters. `data` a list with 3 components: `x` argument `x`. `y` argument `y`. `med` (weighted) mean of the (binned) data.

If `y == NULL`, `bin.den` is called and a `bin.den`-`class` object is returned.

## References

Wand M.P. (1994) Fast Computation of Multivariate Kernel Estimators. Journal of Computational and Graphical Statistics, 3, 433-445.

`data.grid`, `locpol`, `bin.den`, `h.cv`.

## Examples

 ```1 2 3 4 5 6``` ```with(earthquakes, spoints(lon, lat, mag, main = "Earthquake data")) bin <- binning(earthquakes[, c("lon", "lat")], earthquakes\$mag, nbin = c(30,30), set.NA = TRUE) simage(bin, main = "Binning averages") with(earthquakes, points(lon, lat, pch = 20)) ```

### Example output  ``` Package npsp: Nonparametric Spatial Statistics
version 0.5-3 (built on 2016-09-28).