Problems?

There are different kinds of problems you can encounter with these packages. Don't hesitate to contact us if you find bugs in the programs or if you have suggestions for improvements. But we hope you understand that we can't give any support in using the routines or in explaining the methods in general.

Here are some suggestions on problems we know of:

Conceptual problems

Q: What is the meaning of the algorithms? What can they do for me, my time series application?

A: Read the Introduction to the methods. Also read the book:

H. Kantz and T. Schreiber
Nonlinear Time Series Analysis, 2nd edition
Cambridge University Press (2004)

and the specific references given in the documentation.

Usage problems

Q: What do the options mean? How should I set them?

A: Read the documentation for your program and the references given. Re-read the book. If you know what you want to do with the programs, you will probably know how to set the parameters.

Implementational problems

Q: What is the program doing in detail? How is it doing things?

A: Read the Introduction to the methods. Have a look at the source code (don't expect comments or other gimmicks).

Strange results

Q: What do the results mean?

A: Try some simple examples first, using generated data. Check if you installed everything correctly. If the programs work on standard data, but you can't figure out what they do on your experimental time series, it is likely that there is a scope mismatch between the methods and your data. Occasionally, it might be just a formatting or preprocessing problem which should be easy to fix for you. If you have strong reasons to believe that the programs are not doing what they should, contact the author.

Program crashes

Q: What is wrong?

A: Make sure the options are o.k. and the files exist/have the correct permissions. Not much checking is done by the programs. If you believe that you have detected a malfunction, contact the author of the program.

TISEAN home