ami and correlation

Computes and plots average mutual information and correlation for time series data.

AMI computes and plots average mutual information (ami) and correlation of univariate or bivariate time series for different values of time lag.
USAGE:
[amis corrs] = ami(xy,nBins,nLags)

INPUT:
xy: either univariate (x) or bivariate ([x y]) time series data. If bivariate time series are given then x should be independent variable and y should be dependent variable. If univariate time series is given then autocorrelation is calculated instead of cross correlation.

nBins: number of bins for time series data to compute distribution which is required to compute ami. nBins should be either vector of 2 elements (for bivariate) or scalar (univariate).

nLags: number of time lags to compute ami and correlation. Computation is done for lags values of 0:nLags.

OUTPUT:
amis: vector of average mutual information for time lags of 0:nLags

corrs: vector of correlation (or autocorrelation for univariate time seris) for time lags of 0:nLags

EXAMPLES:
xy = rand(1000,2);
nBins = [15 10];
nLags = 25;
[amis corrs]= ami(xy,nBins,nLags);