Addressing Noise and Skewness in Interpretable Health-Condition Assessment by Learning Model Confidence.

Assessing the health condition has a wide range of applications in healthcare, military, aerospace, and industrial fields. Nevertheless, traditional feature-engineered techniques involve manual feature extraction, which are too cumbersome to adapt to the changes caused by the development of sensor network technology. Recently, deep-learning-based methods have achieved initial success in health-condition assessment research, but insufficient considerations for problems such as class skewness, noisy segments, and result interpretability make it difficult to apply them to real-world applications. In this paper, we propose a K-margin-based Interpretable Learning approach for health-condition assessment. In detail, a skewness-aware RCR-Net model is employed to handle problems of class skewness. Furthermore, we present a diagnosis model based on K-margin to automatically handle noisy segments by naturally exploiting expected consistency among the segments associated with each record. Additionally, a knowledge-directed interpretation method is presented to learn domain knowledge-level features automatically without the help of human experts which can be used as an interpretable decision-making basis. Finally, through experimental validation in the field of both medical and aerospace, the proposed method has a better generality and high efficiency with 0.7974 and 0.8005 F1 scores, which outperform all state-of-the-art deep learning methods for health-condition assessment task by 3.30% and 2.99%, respectively.

Opportunities and challenges of deep learning methods for electrocardiogram data: A systematic review.

BACKGROUND: The electrocardiogram (ECG) is one of the most commonly used diagnostic tools in medicine and healthcare. Deep learning methods have achieved promising results on predictive healthcare tasks using ECG signals. OBJECTIVE: This paper presents a systematic review of deep learning methods for ECG data from both modeling and application perspectives. METHODS: We extracted papers that applied deep learning (deep neural network) models to ECG data that were published between January 1st of 2010 and February 29th of 2020 from Google Scholar, PubMed, and the Digital Bibliography & Library Project. We then analyzed each article according to three factors: tasks, models, and data. Finally, we discuss open challenges and unsolved problems in this area. RESULTS: The total number of papers extracted was 191. Among these papers, 108 were published after 2019. Different deep learning architectures have been used in various ECG analytics tasks, such as disease detection/classification, annotation/localization, sleep staging, biometric human identification, and denoising. CONCLUSION: The number of works on deep learning for ECG data has grown explosively in recent years. Such works have achieved accuracy comparable to that of traditional feature-based approaches and ensembles of multiple approaches can achieve even better results. Specifically, we found that a hybrid architecture of a convolutional neural network and recurrent neural network ensemble using expert features yields the best results. However, there are some new challenges and problems related to interpretability, scalability, and efficiency that must be addressed. Furthermore, it is also worth investigating new applications from the perspectives of datasets and methods. SIGNIFICANCE: This paper summarizes existing deep learning research using ECG data from multiple perspectives and highlights existing challenges and problems to identify potential future research directions.

Combining deep neural networks and engineered features for cardiac arrhythmia detection from ECG recordings.

OBJECTIVE: We aim to combine deep neural networks and engineered features (hand-crafted features based on medical domain knowledge) for cardiac arrhythmia detection from short single-lead ECG recordings. APPROACH: We propose a two-stage method named ENCASE for cardiac arrhythmia detection. The first stage is feature extraction and the second stage is classifier building. In the feature extraction stage, we extract both deep features and engineered features. Deep features are obtained by modifying deep neural networks into a deep feature extractor. Engineered features are extracted by summarizing existing approaches into four feature groups. Then, we propose a feature aggregation approach to combine these features. In the classifier building stage, we build multiple gradient boosting decision trees and combine them to get the final detector. MAIN RESULTS: Experiments are performed on the PhysioNet/Computing in Cardiology Challenge 2017 dataset (Clifford et al 2017 Computing in Cardiology vol 44). Using F 1 scores reported on the hidden test set as measurements, ENCASE got 0.9117 on Normal (F 1N ), 0.8128 on Atrial Fibrillation (AF) (F 1A ), 0.7505 on Others (F 1O ), and 0.5671 on Noise (F 1P ). It placed 5th in the Challenge and 8th in the follow-up challenge (ranked by considering the average of Normal, AF, and Others (F 1NAO = 0.825)). When rounding to two decimal places, we were part a three-way tie for 1st place and were part a seven-way tie for 2nd place in the follow-up challenge. Further experiments show that combined features perform better than individual features, and deep features show more importance scores than other features. SIGNIFICANCE: ENCASE can benefit from both feature engineering-based methods and recent deep neural networks. It is flexible and can easily assimilate the ability of new cardiac arrhythmia detection methods.