Recent studies suggested that ST-Elevation Myocardial Infarction (STEMI) can be detected in the ECG relying on machine learning (ML) algorithms. However, most of ML algorithms lack of an interpretability analysis, since they do not provide any justification for their decisions. In this study, we trained a Random Forest (RF) on the Physionet PTB database to automatically detect STEMI patients, considering 12-lead average templates as input. Then, we used the Local Interpretable Model-agnostic Explanations (LIME) method to highlight the input parts that mostly contributed to the detection. LIME interpretations were validated with the anatomical position of the myocardial infarction available within the dataset. Experimental results showed that RF achieved a high test set accuracy (ranging from 0.84 to 0.92). However, LIME identified areas within QRS complexes as the most relevant ones for the RF decision, rather than in the ST segment as expected. Our study suggests that, despite the test set accuracy, ML algorithms for STEMI classification, trained on small or unbalanced/biased populations, may rely on features which are not clinically significant. In this regard, interpretability algorithms like LIME may help in understanding possible pitfalls.
Interpretability Analysis of Machine Learning Algorithms in the Detection of ST-Elevation Myocardial Infarction / M. Bodini, M.W. Rivolta, R. Sassi - In: Computing in Cardiology[s.l] : IEEE Computer Society, 2020. - ISBN 9781728173825. - pp. 1-4 (( Intervento presentato al 47. convegno CinC tenutosi a Rimini nel 2020.
Interpretability Analysis of Machine Learning Algorithms in the Detection of ST-Elevation Myocardial Infarction
M. Bodini
Primo
;M.W. Rivolta;R. SassiUltimo
2020
Abstract
Recent studies suggested that ST-Elevation Myocardial Infarction (STEMI) can be detected in the ECG relying on machine learning (ML) algorithms. However, most of ML algorithms lack of an interpretability analysis, since they do not provide any justification for their decisions. In this study, we trained a Random Forest (RF) on the Physionet PTB database to automatically detect STEMI patients, considering 12-lead average templates as input. Then, we used the Local Interpretable Model-agnostic Explanations (LIME) method to highlight the input parts that mostly contributed to the detection. LIME interpretations were validated with the anatomical position of the myocardial infarction available within the dataset. Experimental results showed that RF achieved a high test set accuracy (ranging from 0.84 to 0.92). However, LIME identified areas within QRS complexes as the most relevant ones for the RF decision, rather than in the ST segment as expected. Our study suggests that, despite the test set accuracy, ML algorithms for STEMI classification, trained on small or unbalanced/biased populations, may rely on features which are not clinically significant. In this regard, interpretability algorithms like LIME may help in understanding possible pitfalls.File | Dimensione | Formato | |
---|---|---|---|
CinC2020-403.pdf
accesso aperto
Tipologia:
Publisher's version/PDF
Dimensione
500.7 kB
Formato
Adobe PDF
|
500.7 kB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.