letsR (version 1.0)

lets.correl: Compute correlogram based on Moran's I

Description

Compute Moran's correlogram of a variable using a distance matrix.

Usage

lets.correl(x, y, z, equidistant=FALSE, plot=TRUE)

Arguments

x

A single variable in vector format or multiple variables in matrix format (as columns).

y

A distance matrix of class matrix.

z

The number of distance classes to use in the correlogram.

equidistant

Logical, if TRUE the classes will be equidistant. If FALSE the classes will have equal number of observations.

plot

Logical, if TRUE the correlogram will be ploted.

Value

• Returns a matrix with the Moran's I Observed value, Standard Deviation and Expected value. Also the p value of the null model, the mean distance between classes and the number of observations.

Examples

var <- runif(100)  # random variable

# Correlated distance matrix
distan <- matrix(runif(1000), ncol=100, nrow=100)
diag(distan) <- 0
distan[lower.tri(distan)] <- distan[upper.tri(distan)]
distan2 <- as.matrix(dist(var))
distan <- (distan)*(distan2)

moran <- lets.correl(var, distan, 5, equidistant=FALSE, plot=TRUE)