國立高雄大學圖資館 |

Language: English

Back

Enhanced EEG Emotion Recognition: A Hybrid Stacked Ensemble Approach With Advanced Computational Visualization /

Record Type:	Electronic resources : Monograph/item
Title/Author:	Enhanced EEG Emotion Recognition: A Hybrid Stacked Ensemble Approach With Advanced Computational Visualization /Kumar Rajmani Bapat.
Author:	Bapat, Kumar Rajmani,
Description:	1 electronic resource (70 pages)
Notes:	Source: Masters Abstracts International, Volume: 86-06.
Notes:	Advisors: Sultornsanee, Sivarit Tony; Lin, Yingzi Committee members: Kamarthi, Sagar; Radhakrishnan, Srinivasan.
Contained By:	Masters Abstracts International86-06.
Subject:	Computer engineering.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31149159
ISBN:	9798346870944

Enhanced EEG Emotion Recognition: A Hybrid Stacked Ensemble Approach With Advanced Computational Visualization /
Bapat, Kumar Rajmani,

Enhanced EEG Emotion Recognition: A Hybrid Stacked Ensemble Approach With Advanced Computational Visualization /Kumar Rajmani Bapat. - 1 electronic resource (70 pages)

Source: Masters Abstracts International, Volume: 86-06.

Brain-computer interfaces (BCIs) hold promise for various applications, with emotion classification from electroencephalography (EEG) being a significant research focus. However, accurately classifying emotions from EEG data remains a challenge due to the complexity and individual variation of emotional signatures in the brain. This paper proposes a novel Hybrid Ensemble Stacking Model (HESM) with network analysis integration to enhance emotion classification accuracy using EEG signals. Our approach leverages a diverse set of base classifiers (Logistic Regression, Support Vector Machines, and a Multi- Layer Perceptron) in a stacking ensemble architecture with dynamic model selection by a meta-learner. Additionally, we utilize a specialized Convolutional Neural Network (CNN) tailored for extracting spatiotemporal EEG patterns relevant to emotions. To further improve our model's performance, we introduce network analysis techniques. We construct functional connectivity networks from EEG time series data using measures such as coherence and phase locking value. We then extract features like centrality measures, clustering coefficients, and path lengths, which capture the topological structure of brain networks associated with emotional states. These network features are integrated into the HESM alongside traditional EEG features, providing richer insights for classification. The proposed methodology is evaluated on an EEG dataset containing positive, negative, and neutral emotional states. We rigorously tune hyperparameters using grid search for both the HESM components and the CNN. Results demonstrate that our hybrid model, augmented with network analysis, leads to significantly improved emotion classification performance compared to standalone machine learning models and CNN-only approaches. Providing a high accuracy with less amount of runtime and reduces the risk of overfitting and is suitable for imbalanced data.This research highlights the potential of combining traditional machine learning, deep learning, and network analysis to decode complex emotional patterns from EEG data for robust BCI applications.

English

ISBN: 9798346870944Subjects--Topical Terms:

212944
Computer engineering.
Subjects--Index Terms:

EEG

Enhanced EEG Emotion Recognition: A Hybrid Stacked Ensemble Approach With Advanced Computational Visualization /
LDR:03734nmm a22004573i 4500 001 676299
005 20250420214338.5
006 m o d
007 cr|nu||||||||
008 250901s2024 miu||||||m |||||||eng d
020 $a 9798346870944
035 $a (MiAaPQD)AAI31149159
035 $a AAI31149159
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Bapat, Kumar Rajmani, $e author. $3 990058
245 1 0 $a Enhanced EEG Emotion Recognition: A Hybrid Stacked Ensemble Approach With Advanced Computational Visualization / $c Kumar Rajmani Bapat.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2024
300 $a 1 electronic resource (70 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Masters Abstracts International, Volume: 86-06.
500 $a Advisors: Sultornsanee, Sivarit Tony; Lin, Yingzi Committee members: Kamarthi, Sagar; Radhakrishnan, Srinivasan.
502 $b M.S. $c Northeastern University $d 2024.
520 $a Brain-computer interfaces (BCIs) hold promise for various applications, with emotion classification from electroencephalography (EEG) being a significant research focus. However, accurately classifying emotions from EEG data remains a challenge due to the complexity and individual variation of emotional signatures in the brain. This paper proposes a novel Hybrid Ensemble Stacking Model (HESM) with network analysis integration to enhance emotion classification accuracy using EEG signals. Our approach leverages a diverse set of base classifiers (Logistic Regression, Support Vector Machines, and a Multi- Layer Perceptron) in a stacking ensemble architecture with dynamic model selection by a meta-learner. Additionally, we utilize a specialized Convolutional Neural Network (CNN) tailored for extracting spatiotemporal EEG patterns relevant to emotions. To further improve our model's performance, we introduce network analysis techniques. We construct functional connectivity networks from EEG time series data using measures such as coherence and phase locking value. We then extract features like centrality measures, clustering coefficients, and path lengths, which capture the topological structure of brain networks associated with emotional states. These network features are integrated into the HESM alongside traditional EEG features, providing richer insights for classification. The proposed methodology is evaluated on an EEG dataset containing positive, negative, and neutral emotional states. We rigorously tune hyperparameters using grid search for both the HESM components and the CNN. Results demonstrate that our hybrid model, augmented with network analysis, leads to significantly improved emotion classification performance compared to standalone machine learning models and CNN-only approaches. Providing a high accuracy with less amount of runtime and reduces the risk of overfitting and is suitable for imbalanced data.This research highlights the potential of combining traditional machine learning, deep learning, and network analysis to decode complex emotional patterns from EEG data for robust BCI applications.
546 $a English
590 $a School code: 0160
650 4 $a Computer engineering. $3 212944
650 4 $a Computer science. $3 199325
650 4 $a Information technology. $3 184390
653 $a EEG
653 $a Linear SVC
653 $a Logistic regression
653 $a Multi-Layer Perceptron
653 $a Support Vector Machines
690 $a 0464
690 $a 0984
690 $a 0489
690 $a 0800
710 2 $a Northeastern University. $b Mechanical and Industrial Engineering. $3 603360
720 1 $a Sultornsanee, Sivarit Tony $e degree supervisor.
720 1 $a Lin, Yingzi $e degree supervisor.
773 0 $t Masters Abstracts International $g 86-06.
790 $a 0160
791 $a M.S.
792 $a 2024
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31149159