WEKA and event related potentials
-
1
University of West Bohemia, Czechia
Introduction: The process of preprocessing, processing and classification of EEG/ERP signals is a very complex task, for which various software tools are used. However, some methods are used frequently in different stages of the task solution. Thus we decided to use advantages of software WEKA, which represents very powerful framework with strict rules. It also contains a great amount of methods such as wavelet transformation, Bayes classificators in many versions, artificial neural networks, decision trees etc. However WEKA software does not include some algorithms and data structures important for EEG/ERP research and signal processing.
At first we designed a new way how WEKA could handle relational data, because complex data obtained by EEG devices need to be organized in 1:N relationship at least (see Figure 1). This proposal includes design of a new data file format and also new methods for data preprocessing and preparation to enable their usage within standard methods included in WEKA. This part is based on SQL queries; it means that it is not necessary to store data only into files. WEKA and our plug-in can read data directly from database.
The second part of our task was to implement fundamental signal preprocessing techniques such as convolution, correlation and others. It is very important to use e.g. numeric filters to remove noise. We intend to include methods like matching pursuit or ICA, which are handy for artifacts rejection. We also added artificial neural network ART-2 to WEKA.
The third part deals with visualization of EEG signals and database tables. We plan to implement new components for signal visualization and also for ERA diagrams visualization.
How does it work?
All our methods and components are packed in .jar file. This file is included in WEKA class path. WEKA itself reads our classes and integrates them. This feature ensures that our GUI and methods are accessible through WEKA GUI (Explorer, Knowledge Flow). Thus there is no need to use other software tools for signal processing and pre-processing.
Conclusion: We work on a very powerful tool for EEG/ERP signal processing and preprocessing. The plug-in for WEKA and its source code will be soon accessible through SVN repository. Our aim is also to encourage all possible participants to help us add other methods, which can be used by other researchers as well. WEKA, enriched with our new plug-in, is also suitable for education of new students and beginners in signal processing, data mining and EEG/ERP domain.
References
1. Ian H. Witten & Eibe Frank, Data Mining: Practical Machine Learning Tools and Techniques, Morgan Kaufman Publisher, ISBN:978-0-12-088407-0
2. S. W. Smith, The Scientist and Engineer’s Guide to Digital Signal Processing, California Technical Publishing, California, 1997, URL:http://www.dspguide.com/.
3. L. Fausett, Fundamentals of Neural Networks, Prentice-Hall, New Jersey, 1994
4. http://www.cs.waikato.ac.nz/ml/weka/
5. http://www.hakank.org/weka/
6. http://weka.sourceforge.net/doc/
Conference:
Neuroinformatics 2008, Stockholm, Sweden, 7 Sep - 9 Sep, 2008.
Presentation Type:
Poster Presentation
Topic:
General Neuroinformatics
Citation:
Ciniburk
J and
Souhrada
V
(2008). WEKA and event related potentials.
Front. Neuroinform.
Conference Abstract:
Neuroinformatics 2008.
doi: 10.3389/conf.neuro.11.2008.01.003
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
25 Jul 2008;
Published Online:
25 Jul 2008.
*
Correspondence:
Jindrich Ciniburk, University of West Bohemia, Pilsen, Czechia, ciniburk@kiv.zcu.cz