Mutual Information Iterated Local Search: A Wrapper-Filter Hybrid for Feature Selection in Brain Computer Interfaces

Adair, Jason; Brownlee, Alexander E. I.; Ochoa, Gabriela

doi:10.1007/978-3-319-77538-8_5

Cited by 3 publications

(2 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Jason Adair et al in their experiments used feature extractors like entropy, mutual information and maximum relevance minimal redundancy to extract the essential features and supplied it to classifiers like KNN and SVM to achieve the accuracy of 60-73% and other notable parameters that they deduced were the number of features and CVE [17]. Indu Dokare et al, in their research, exercised the wavelet transform methodology to extract features after pre-processing the data and decompose the signal into 6 wavelet coefficients.…”

Section: Comparison To Other Related Research Workmentioning

confidence: 99%

Identifying EEG Binary Limb Motor Imagery Movements using Thick Data Analytics

Trikha¹,

Fiaidhi²,

Mohammed³

2020

APJCRI

View full text Add to dashboard Cite

Electroencephalography (EEG) is non-invasive technology that is widely used to record brain signals in brain computer interfacing (BCI) systems to control, motor imagery, in which movements signals occurring in limbs can control some services. Researchers have proposed numerous classification schemes of these motor imagery to incorporate it with various neurorehabilitation, neuroprosthetics and gaming applications. However, the existing classification schemes face the performance degradation caused by motor-imagery EEG signals with low signal to noise ratio. The paper's main objective is to use possible thick data analytics techniques to classify effectively the motor imagery EEG signals. Our attempt start with notable classifiers including Decision Trees, Extra Trees, Naive Bayes, Random Forest and SVM and move later to enhance classifications using variety of ensemble learning techniques including Bagging, Adaboost and Stacking. More techniques has been applied on the results of the ensemble learring to eliminate classification noise and supply more relevant features such as substituting outliers with mean value and exercising band-pass filter and Common Spatial Pattern (CSP). The thick data methods has been validated on a public dataset rendered by BCI competition II dataset III and was found to produce better classification performance metric which included performance metric parameters like accuracy, specificity, sensitivity, precision and recall when confronted with the existing work, thus projecting the usefulness of motor imagery BCI. The analytics is inclusive of Area Under the Curve (AUC) score and Mathews Correlation Coefficient (MCC) score to display an impactful analysis.

show abstract

Section: Comparison To Other Related Research Workmentioning

confidence: 99%

Identifying EEG Binary Limb Motor Imagery Movements using Thick Data Analytics

Trikha¹,

Fiaidhi²,

Mohammed³

2020

APJCRI

View full text Add to dashboard Cite

show abstract

“…Several global and local search algorithms have been deployed for optimization purposes. New filterwrapper hybrid based were invented by Adair, Brownlee and Ochoa (2018) and Rodriguez-Galiano, Luque-Espinar, Chica-Olmo and Mendes (2018) . The proposed method has provided more reliable solutions, where the solutions are more able to generalize unseen data.…”

Section: Related Workmentioning

confidence: 99%

Optimization of Attribute Selection Model Using Bio-Inspired Algorithms

Basir

Yusof

Hussin

2019

JICT

View full text Add to dashboard Cite

Attribute selection which is also known as feature selection is an essential process that is relevant to predictive analysis. To date, various feature selection algorithms have been introduced, nevertheless they all work independently. Hence, reducing the consistency of the accuracy rate. The aim of this paper is to investigate the use of bio-inspired search algorithms in producing optimal attribute set. This is achieved in two stages; 1) create attribute selection models by combining search method and feature selection algorithms, and 2) determine an optimized attribute set by employing bio-inspired algorithms. Classification performance of the produced attribute set is analyzed based on accuracy and number of selected attributes. Experimental results conducted on six (6) public real datasets reveal that the feature selection model with the implementation of bio-inspired search algorithm consistently performs good classification (i.e higher accuracy with fewer numbers of attributes) on the selected data set. Such a finding indicates that bio-inspired algorithms can contribute in identifying the few most important features to be used in data mining model construction.

show abstract

A Survey on Feature Selection, Classification, and Optimization Techniques for EEG-Based Brain–Computer Interface

Subramanian,

Daniel

2024

Lecture Notes in Networks and Systems

View full text Add to dashboard Cite

Mutual Information Iterated Local Search: A Wrapper-Filter Hybrid for Feature Selection in Brain Computer Interfaces

Cited by 3 publications

References 25 publications

Identifying EEG Binary Limb Motor Imagery Movements using Thick Data Analytics

Identifying EEG Binary Limb Motor Imagery Movements using Thick Data Analytics

Optimization of Attribute Selection Model Using Bio-Inspired Algorithms

A Survey on Feature Selection, Classification, and Optimization Techniques for EEG-Based Brain–Computer Interface

Contact Info

Product

Resources

About