Explainable artificial intelligence for heart rate variability in ECG signal

Sanjana, K; Sowmya, V.; Gopalakrishnan, E. A.; Soman, K. P.

doi:10.1049/htl.2020.0033

Cited by 24 publications

(4 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the end-users of the models, including physicians and clinical experts, understanding the internal workings and decision logic of the models is crucial [27]. This opacity can lead to a decrease in trust in the model's predictions by medical professionals, particularly in medical diagnostic situations where a detailed explanation of the model's behavior is required [7].…”

Section: Interpretability Analysismentioning

confidence: 99%

Section: Comparison With Existing Workmentioning

confidence: 99%

“…In the research of ECG signal classification, deep learning models such as CNNs have achieved remarkable results [28,29], but the opacity of their decision-making processes can lead to a decrease in trust in clinical applications [7]. In the domain of ECG signal classification, SHAP value analysis has proven to be a powerful tool for identifying and interpreting key features that affect model predictions [7]. Existing studies have shown that SHAP values play a significant role in recognizing crucial frequency features in ECG signals, predicting ST-segment elevations for cardiac events, and offering other diagnostic insights [5,6,[25][26][27].…”

Section: Comparison With Existing Workmentioning

confidence: 99%

“…Against this backdrop, deep learning techniques, particularly Convolutional Neural Networks (CNNs), have gradually become cutting-edge technology in this field due to their outstanding data-driven feature extraction and classification capabilities [4]. However, CNN models are often regarded as "black boxes," lacking transparency in their decision-making processes [5,6], and coupled with the issue of class imbalance in ECG datasets, these factors have limited their application in clinical practice [7]. To overcome these challenges, this study is dedicated to developing a CNN model that can not only handle the task of automatic 2 classification of imbalanced multi-class ECG signals but also possesses a high degree of interpretability to meet the professional needs of the medical diagnostic field.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Convolutional Neural Network-Based ECG Signal Classification Model: A Study on Addressing Class Imbalance and Enhancing Model Interpretability

Wang,

Zheng,

Yang

et al. 2024

Preprint

View full text Add to dashboard Cite

Convolutional Neural Networks (CNNs) are often criticized for their lack of transparency, acting as 'black boxes' in decision-making, a challenge compounded by class imbalance in ECG datasets, which limits their clinical application. This study introduces a CNN-based ECG signal classification model that enhances interpretability and addresses class imbalance through the Synthetic Minority Over-sampling Technique (SMOTE). The model also integrates Uniform Manifold Approximation and Projection (UMAP) for dimensionality reduction and SHAP value analysis, facilitating the visualization of decision boundaries and the assessment of feature contributions. Our evaluation using the MIT-BIH Arrhythmia Database highlights the model's high performance, with accuracy and precision nearing 1.00 for Normal ECG (NOR), Left Bundle Branch Block (LBBB), Right Bundle Branch Block (RBBB), and Ventricular Premature Beat (PVC) in the six-class task. In the ten-class task, the model demonstrated robustness, particularly with an accuracy of 0.9846, precision of 0.9783, recall of 0.9736, and F1 score of 0.9760 in Pacemaker Fusion Beat (PFHB), supported by an AUC of 0.9999 and AP of 0.9885. These results underscore the model's efficacy in cardiac rhythm recognition and resilience to class imbalance. Future research will explore sophisticated model architectures and feature extraction methods to enhance the model's generalization and clinical applicability for early heart disease diagnosis and personalized treatment.

show abstract

Section: Interpretability Analysismentioning

confidence: 99%