An Ensemble Pattern Classification System Based on Multitree Genetic Programming for Improving Intension Pattern Recognition Using Brain Computer Interaction

Abstract

Ensemble learning is one of the successful methods to construct a classification system. Many researchers have been interested in the method for improving the classification accuracy. In this paper, we proposed an ensemble classification system based on multitree genetic programming for intension pattern recognition using BCI. The multitree genetic programming mechanism is designed to increase the diversity of each ensemble classifier. Also, the proposed system uses an evaluation method based on boosting and performs the parallel learning and the interaction by multitree. Finally, the system is validated by the comparison experiments with existing algorithms. © Springer-Verlag Berlin Heidelberg 2014.