plotClusters: Plot the field samples as clusters

In USGS-R/GcClust: Clustering geochemical data with HMC sampling

Plot the field samples as clusters

Description

Plot the locations of the field samples on a previously-plotted map. The attributes of each location symbol (for example, color) indicate the cluster to which the field sample belongs. That is, the attributes indicate the conditional probability that the field sample is associated with a particular pdf in the finite mixture model.

Usage

plotClusters(gcData, condProbs1, probIntervals = c(0, 0.1, 0.5, 0.9, 1),
  symbolIndices = c(16, 16, 16, 16), symbolSizes = c(1/3, 1/3, 1/3, 1/3),
  symbolColors = c("red", "yellow", "green", "blue"))

Arguments

gcData	List containing the geochemical and related data. This container is described in the package documentation.
condProbs1	A matrix containing the Monte Carlo samples of the conditional probabilities. This matrix is returned by function calcCondProbs1, for which the documentation includes a complete description of container condProbs1.
probIntervals	Vector containing intervals of conditional probability. All field samples within an given interval are plotted the same way.
symbolIndices	Vector containing the indices of the plotting symbols for the conditional probability intervals.
symbolSizes	Vector containing the relative sizes of the plotting symbols for the conditional probability intervals.
symbolColors	Vector containing the colors of the plotting symbols for the conditional probability intervals.

Details

The conditional probabilities indicate the extend to which the field samples are associated with the first pdf in the finite mixture model. The conditional probabilities in container condProbs1 are Monte Carlo samples, and their medians are used to assign plotting attributes for the field samples.

The plotting attributes are specified by arguments probIntervals, symbolIndices, symbolSizes, and symbolColors. To understanding the specification of these attributes, consider their default values, which pertain to four probability intervals.

Vector probIntervals has elements 0, 0.1, 0.5, 0.9, and 1. These five elements specify four probability intervals: [0,0.1], [0.1,0.5], [0.5,0.9] and [0.9,1]. Notice that the first and last elements of probIntervals are 0 and 1 respectively. The probability intervals are used to classify the field samples based upon their associated conditional probabilities:

• If the conditional probability of a field sample is within the interval [0,0.1], then the field sample is classified as "strongly associated with pdf 2" and is assigned the color red. This color is consistent with the colors used in functions plotStdCompMeans and plotCompMeans.

• If the conditional probability of a field sample is within the interval [0.1,0.5], then the field sample is classified as "moderately associated with pdf 2" and is assigned the color yellow.

• If the conditional probability of a field sample is within the interval [0.5,0.9], then the field sample is classified as "moderately associated with pdf 1" and is assigned the color green.

• If the conditional probability of a field sample is within the interval [0.9,1], then the field sample is classified as "strongly associated with pdf 1" and is assigned the color blue. This color is consistent with the colors used in functions plotStdCompMeans and plotCompMeans.

Because, in this explanation, vector probIntervals specifies four probability intervals, arguments for vectors symbolIndices, symbolSizes, and symbolColors must have four elements. The symbol indices, symbol sizes, and symbol colors are described in Murrell (2006, p. 55-56, 68, 69).

This function adds symbols to a map that has already been plotted.

References

Murrell, P., 2006, R graphics: Chapman & Hall / CRC.

Examples

## Not run: 
map('state', fill = TRUE, col = "gray60", border = "white")
plotClusters(concentrationData, condProbs1)

## End(Not run)

USGS-R/GcClust documentation built on April 17, 2023, 8:08 p.m.