Authors:
Mario Michael Krell
1
;
Nils Wilshusen
1
;
Andrei Cristian Ignat
2
and
Su Kyoung Kim
3
Affiliations:
1
University of Bremen, Germany
;
2
Robotics Innovation Center, German Research Center for Artificial Intelligence GmbH and UC Santa Cruz, Germany
;
3
Robotics Innovation Center and German Research Center for Artificial Intelligence GmbH, Germany
Keyword(s):
Support Vector Machine, Online Learning, Brain Computer Interface, Electroencephalogram, Incremental/Decremental Learning.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Biomedical Engineering
;
Biomedical Instruments and Devices
;
Biomedical Signal Processing
;
Brain-Computer Interfaces
;
Data Manipulation
;
Devices
;
EMG Signal Processing and Applications
;
Health Engineering and Technology Applications
;
Human-Computer Interaction
;
Methodologies and Methods
;
Mobile and Embedded Devices
;
Neural Rehabilitation
;
Neural Signal Processing
;
Neurocomputing
;
NeuroSensing and Diagnosis
;
Neurotechnology, Electronics and Informatics
;
Pattern Recognition
;
Physiological Computing Systems
;
Robotic Assisted Therapy
;
Sensor Networks
;
Soft Computing
Abstract:
It is often the case that practical applications of support vector machines (SVMs) require the capability to perform online learning under limited availability of computational resources. Enabling SVMs for online learning can be done through several strategies. One group thereof manipulates the training data and limits its size. We aim to summarize these existing approaches and compare them, firstly, on several synthetic datasets with different shifts and, secondly, on electroencephalographic (EEG) data. During the manipulation, class imbalance can occur across the training data and it might even happen that all samples of one class are removed. In order to deal with this potential issue, we suggest and compare three balancing criteria.