High performance GA-LDA feature selection model for Brain Computer Interface data

Sahar Alwadei, Mohamed Dahab, Mahmoud Kamel


High-Performance Computing (HPC) proved notable performance enhancements in many different applications. The research is considering the Brain-Computer Interface (BCI) data, precisely the P300 based system. BCI is a system that provides a direct communication control channel between the brain and the external world, but its data processing is exceedingly time-consuming. That system consists of many components where Feature Selection is a primary key to its performance. There is a need for search algorithms and classifiers to build a feature selection model using the wrapper approach. Hence, they are the concern of this research where the Genetic Algorithm (GA) implemented as a search algorithm and Linear Discriminant Analysis (LDA) as a classifier. Thus, the model for features selection formed as GA-LDA. The Evolutionary Algorithm (EA) used, GA, estimates an optimal solution and saves an enormous amount of time more than other algorithms such as brute force search. Furthermore, HPC techniques had implemented since the computational power was one of the main obstacles besides the problem's size causing an extensive processing time. This model saves 98.1% of the original time consumed while using common computing facilities. It also maintains an accuracy rate of 72.5% selecting 46.2% of the original features only.

Full Text:



Al-ani T. and Trad D., (2010.), Signal Processing and Classification Approaches for Brain-computer Interface. INTECH Open Access Publisher.

Alharbi, A. N., and Dahab, M., (2018) Comparative Study on Fast Feature Selection. International Journal of Information Technology and Language Studies.

Alwadei, Sahar & Dahab, Mohamed & Kamel, Mahmod, (2017) A Feature Selection Model based on High-Performance Computing (HPC) Techniques. International Journal of Computer Applications. 180. 11-16.

Atom, Yanina, Gareis, Iván, Gentiletti, Gerardo, Acevedo, Rubén, and Rufiner, Leonardo, (2010), Genetic Feature Selection to Optimally Detect P300 in Brain-Computer Interfaces, 32nd Annual International Conference of the IEEE EMBS, Buenos Aires, Argentina.

Blankertz, B., Tangermann, M., Vidaurre, C., Dickhaus, T., Sannelli, C., Popescu, F., Fazli, S., Danóczy, M., Curio, G., and Müller, Klaus-Robert, (2010), Detecting Mental States by Machine Learning Techniques: The Berlin Brain-Computer Interface, B. Graimann et al. (eds.), Brain-Computer Interfaces, The Frontiers Collection, Springer-Verlag Berlin Heidelberg.

Borikar, Shweta Sanjay, Snehal Ramesh Kochre, and Yugandhara D. Zade. (2014), Brain-Computer Interface. National Conference on Engineering Trends in Medical Science (NCETMS).

Cabrera, Alvaro R. Feature Extraction and Classification for Brain-Computer Interfaces. (2009), Ph.D. Thesis, Brain-Computer Interface Laboratory Center for Sensory-Motor Interaction (SMI), Department of Health Science and Technology Aalborg University, Denmark.

Croux, C., Filzmoser P., and Joossens K., (2008), Classification efficiencies for robust linear discriminant analysis, Statistica Sinica, 18(2): p. 581-599.

Cunningham, Pádraig. (2008), Dimension reduction. Machine learning techniques for multimedia. Springer Berlin Heidelberg, 91-112.

Datasets from UCI. SGI, Silicon Graphics International Corp, Retrieved 22 October 2017 from https://www.sgi.com/tech/mlc/db/

Elshout J.and Molina G. G., (2009), Review of brain-computer interfaces based on the p300 evoked potential," tech. rep., Koninklijke Philips Electronics.

Fodor, Isola K. (2002), A survey of dimension reduction techniques.

Fujitsu Supports King Abdul-Aziz University Research Capabilities with New Supercomputing System. Press release on June 01, 2015. King Abdul-Aziz University, Fujitsu Limited. Jeddah and Tokyo,

Hoffmann, Ulrich, Vesin, Jean-Marc, Ebrahimi, Touradj, and Diserens, Karin, (2008), An efficient P300-based brain-computer interface for disabled subjects, Journal of Neuroscience Methods 167, pp. 115–125.

John, Kohavi, and Pfleger, Irrelevant features and the subset selection problem. Machine Learning: Proceedings of the Eleventh International Conference, 1994, Retrieved 22 October 2017 from http://robotics.stanford.edu/~ronnyk . Last access: 10/22/2017.

K Y Lee, M.A. El-Sharkawi, (2008), Modern Heuristic Optimization Techniques, IEEE press and Wiley – InterScience, New Jersey.

Kachitvichyanukul, Voratas. (2012), Comparison of Three Evolutionary Algorithms. Industrial Engineering & Management Systems 11.3 215-223.‏

Kamel, Mahmoud I., and Anas A. Hadi. (2014), Improving P300 Based Speller by Feature Selection. Journal of Medical Imaging and Health Informatics 4.4: 469-487.

Koprinska, Irena. (2010), Feature selection for brain-computer interfaces. New frontiers in applied data mining. Springer Berlin Heidelberg, 106-117.

Lan, Tian. (2011), Feature extraction feature selection and dimensionality reduction techniques for brain-computer interface. Doctor of Philosophy in Electrical Engineering examined and approved thesis. Oregon Health & Science University, OHSU Digital Commons, Scholar Archive, Paper 706.

Luck, Steven J. (2005), An Introduction to the Event-Related Potential Technique Cognitive Neuroscience, Cambridge, Mass. MIT Press.

Rody P S Oldenhuis, (2010), Trajectory Optimization of a mission to the Solar Bow shock and minor planets," MSc thesis report, Delft University of Technology, Netherlands.

Smith, Raymond C. (2004), Electroencephalograph based Brain-Computer Interfaces, Master Thesis, University College Dublin, Dublin, Ireland.

Top 500, The List retrieved 22 October 2017 from http://www.top500.org/site/50585.

Umbarkar, A. J., M. S. Joshi, and P. D. Sheth. OpenMP Dual Population Genetic Algorithm for Solving Constrained Optimization Problems. International Journal of Information Engineering and Electronic Business (IJIEEB) 7.1: 59.

Wolpaw, Jonathan R., et al. (2002), Brain-computer interfaces for communication and control. Clinical neurophysiology 113.6.‏ ‏‏

Xin-She Yang, (2010), Engineering Optimization – An Introduction to Metaheuristic Applications. John Wiley & Sons, Hoboken, New Jersey.