What is the clinical evidence from this study?

Study design: Other. Population: Emotion recognition (n=48). Intervention: Machine learning algorithms with feature selection (EFS and Pearson's correlation) vs. Machine learning algorithms without feature selection. Primary outcome: Classification accuracy of emotional state.

November 7, 2022Open Access

Evaluation of Machine Learning Algorithms for Emotions Recognition using Electrocardiogram

Key Result

The Extra Tree classifier with Pearson's correlation feature selection achieved 100% accuracy for emotion recognition on the SWELL dataset, outperforming other machine learning models.

Structured PICO

Population

48 participants from two public datasets (DREAMER and SWELL) whose ECG data were used to evaluate machine learning algorithms for emotion recognition.

Intervention

Seven machine learning models (MLP, SVM, DT, GBDT, Logistic Regression, Adaboost, Extra Tree) trained on ECG-derived heart rate variability (HRV) features using exhaustive feature selection (EFS) and Pearson’s correlation.

Comparator

Models trained without feature selection, and comparison among the different machine learning models.

Outcome

Accuracy of emotion state classification (valence, arousal, dominance for DREAMER; stress vs no stress for SWELL).

Machine learning models, particularly Extra Tree with feature selection, can accurately recognize human emotional states using single-modality ECG signals.

Main Result

Absolute Event Rate: 100% vs 97.4%

Limitations

Small dataset size (especially DREAMER) limiting the efficacy of learning algorithms
Inability to effectively utilize deep learning due to the lack of a large dataset

Abstract

In recent studies, researchers have focused on using various modalities to recognize emotions for different applications. A major challenge is identifying emotions correctly with only electrocardiograms (ECG) as the modality. The main objective is to reduce costs by using single-modality ECG signals to predict human emotional states. This paper presents an emotion recognition approach utilizing the heart rate variability features obtained from ECG with feature selection techniques (exhaustive feature selection (EFS) and Pearson’s correlation) to train the classification models. Seven machine learning (ML) models: multi-layer perceptrons (MLP), Support Vector Machine (SVM), Decision Tree (DT), Gradient Boosting Decision Tree (GBDT), Logistic Regression, Adaboost and Extra Tree classifier are used to classify emotional state. Two public datasets, DREAMER and SWELL are used for evaluation. The results show that no particular ML works best for all data. For DREAMER with EFS, the best models to predict valence, arousal, and dominance are Extra Tree (74.6%), MLP and DT (74.6%), and GBDT and DT (69.8%), respectively. Extra tree with Pearson’s correlation are the best method for the ECG SWELL dataset and provide 100% accuracy. The usage of Extra tree classifier and feature selection technique contributes to the improvement of the model accuracy. Moreover, the Friedman test proved that ET is as good as other classification models for predicting human emotional state and ranks highest. Doi: 10.28991/ESJ-2023-07-01-011 Full Text: PDF

问 AI

Bookmark

View Full Paper