Improved Transductive Support Vector Machine for a Small Labelled Set in Motor Imagery-Based Brain-Computer Interface.

Yilu Xu,Jizhong Liu,Fumin Guo,Jing Hua,Xin Huang,Ronghua Hu,Hua Zhang

doi:10.1155/2019/2087132

Abstract

Long and tedious calibration time hinders the development of motor imagery- (MI-) based brain-computer interface (BCI). To tackle this problem, we use a limited labelled set and a relatively large unlabelled set from the same subject for training based on the transductive support vector machine (TSVM) framework. We first introduce an improved TSVM (ITSVM) method, in which a comprehensive feature of each sample consists of its common spatial patterns (CSP) feature and its geometric feature. Moreover, we use the concave-convex procedure (CCCP) to solve the optimization problem of TSVM under a new balancing constraint that can address the unknown distribution of the unlabelled set by considering various possible distributions. In addition, we propose an improved self-training TSVM (IST-TSVM) method that can iteratively perform CSP feature extraction and ITSVM classification using an expanded labelled set. Extensive experimental results on dataset IV-a from BCI competition III and dataset II-a from BCI competition IV show that our algorithms outperform the other competing algorithms, where the sizes and distributions of the labelled sets are variable. In particular, IST-TSVM provides average accuracies of 63.25% and 69.43% with the abovementioned two datasets, respectively, where only four positive labelled samples and sixteen negative labelled samples are used. Therefore, our algorithms can provide an alternative way to reduce the calibration time.

Highlights

A brain-computer interface (BCI) system can allow people to communicate directly with electronic equipment using their brain activity and without using their peripheral nerves and muscles [1]
We construct the variation of a weighted graph as proposed by Chapelle [25] in order to explore the potential distribution of all samples in a semisupervised way. en, we introduce a comprehensive feature for each sample, which consists of its common spatial patterns (CSP) feature and its geometric feature
Feature learning is critical for the BCI system. us, we develop an improved self-training transductive support vector machine (TSVM) (IST-TSVM) method that can execute CSP and our proposed improved TSVM (ITSVM) method jointly and iteratively. e contributions of our work are summarized as follows: (1) We propose an ITSVM method that can maximize the margin between different clusters and provide different views of all samples based on their CSP and geometric features

Summary

Introduction

A brain-computer interface (BCI) system can allow people to communicate directly with electronic equipment using their brain activity and without using their peripheral nerves and muscles [1]. In a noninvasive BCI system, electroencephalogram (EEG) signals are used to measure brain activity due to their safety and convenience [2]. An MI-based BCI system is suitable for use in military, entertainment, and rehabilitation engineering systems. Due to the inherent nonstationarity of EEG signals, long and tedious calibration time is one of the key issues preventing broad use of MI-based BCI [4, 5]. Reducing the calibration time without loss of accuracy is a major challenge. To solve this problem, semisupervised learning (SSL) classifiers can use a small labelled set and a relatively large unlabelled set from the same subject for training

Methods

Results

Discussion

Conclusion