% Description: Testing for Henon maps
% DocumentationOf: henon_map.m
% Non-linear time series analysis
% Exercise 3.1
clear all;
close all;
pointSet = tim.henon_map();
% The first minimum of auto mutual information fails to
% predict a good embedding lag for this data set.
% The reason is that the Henon map lacks continuity, an
% assumption of the automated method. The best embedding lag
% is empirically found to be 1 (just try different choices).
% I suspect this is a general property of an iterated
% function; every step is important.
% The auto mutual information goes to zero almost immediately;
% again I suspect this is a general property of an iterated function.
tim.embedding_plot(pointSet, 'tDelta', 1);