Contents: Introduction | Utilities | Surrogate Data | Prediction | Delta-Epsilon | Toy Problems | Download | Matlab | Montreal Summer School | Related Sites

Matlab Software for Nonlinear Time-Series Analysis

Danny Kaplan
Dept. of Mathematics
Macalester College
1600 Grand Ave.
St. Paul, MN 55105
kaplan@macalester.edu

This site provides access to a small collection of software implementing several techniques of nonlinear dynamics time series analysis. In addition to providing the software, which is written in the Matlab language, I give some examples of its use.

Most of the software was initially written for the Montreal Summer School on Nonlinear Dynamics in Biology and Medicine, where it is used in the time-series labs. However, I have had requests for the software from researchers in diverse fields.

The software is a complete package for nonlinear-dynamics time series analysis. By "complete" I mean simply that the only additional things you need to use the software are 1) some data and 2) matlab. The techniques included emphasize nonlinear prediction and surrogate data, as well as my own delta-epsilon technique. Many important techniques are not represented, for instance algorithms for estimating lyapunov exponents. This selection reflects my own research interests and pedagogical style, and is in no way a comment on those missing methods.

Before using this software for actual research problems, try it out on toy problems. This will help to make sure that the software is working right on your system, and will also help ensure that you are using the software in a manner consistent with its design

Legal stuff: I make no warrantee of any sort about this software and how it functions. No commercial use may be made of this software without permission from me. If you download the software, you will be asked to fill in a short form, giving your name, address etc. The main purpose of this is so that I know who might be using the software so that I can contact them as bugs are discovered.