Systematic Review on Machine-Learning Algorithms Used in Wearable-Based eHealth Data Analysis

Abstract

In this digitized world, data has become an integral part in any domain, including healthcare. The healthcare industry produces a huge amount of digital data, by utilizing information from all sources of healthcare, including the patients' demographics, medications, vital signs, physician's observations, laboratory data, billing data, data from various wearable sensors, etc. With the rapid growth of the wireless technology applications, there has also been a significant increase in the digital health data. New medical discoveries and new eHealth-related technologies, such as mobile apps, novel sensors, and wearable technology have contributed as important data sources for healthcare data. Nowadays, there is a huge potential to improve the healthcare quality and customer satisfaction with the help of machine learning (ML) algorithms applied on time-domain and frequency-domain healthcare data obtained from wearables and sensors. This systematic literature review examines in depth how health data from sensors can be processed and analyzed using ML techniques. The review focuses on the following diseases for obtaining the eHealth data: diabetes mellitus type 1 and type 2, hypertension and hypotension, atrial fibrillation, bradykinesia, dyskinesia, and fever related diseases. The data for the systematic literature review was collected from four databases, Medline, Proquest, Scopus, and Web of Science. We selected 67 studies for the final in-depth review out of the initial pre-selected papers. Our study identified that the major part of eHealth data is obtained from the sensors such as accelerometer, gyroscopes, ECG (Electrocardiogram), EEG (Electroencephalogram ) monitors, and blood glucose sensors. This study also examines the different feature types, feature extraction methods, and ML algorithms used for eHealth data analysis. Our review also shows that neural network (NN) algorithms and support vector machines (SVM) have shown so far the best performance for analyzing the healthcare data among other ML algorithms studied in the literature.

Managing Perceived Loneliness and Social-Isolation Levels for Older Adults: A Survey with Focus on Wearables-Based Solutions

“…In terms of what kind of ML algorithms are best suitable to work with such data, a systematic literature study conducted in [ 140 ] for ML algorithms used for eHealth applications showed the results depicted in the pie chart in Figure 6 . These results from Figure 6 are based on a systematic review of 67 scientific articles.…”

“…Other NN-based algorithms, such as Convolutional Neural Network (CNN) or Recurrent Neural Network (RNN), were mainly used for image classification but not widely encountered in the context of eHealth data. DNN and multilayer perceptron showed better performance for classification tasks for time domain and frequency domain values than other studied algorithms in [ 140 ] and references therein. Ensemble learning techniques have also been used in of studies.…”

“…6. Distribution of the use of ML algorithms for eHealth applications, based on data from [140] In terms of what kind of ML algorithms are best suitable to work with such data, a systematic 958 literature study conducted in [140] for ML algorithms used for eHealth applications showed the 959 results depicted in the pie chart in Fig. 6.…”

“…From the discussion, it can be inferred that for analyzing the features from continuous sensor 983 data, NN, SVM, and tree-based algorithms such as random forests shows better performance 984 than other ML algorithms, and we are thus recommending one of these three algorithm types 985 (namely SVM, NN, or random forest) to be used for loneliness monitoring and prediction.A 986 Figure 6. Distribution of the use of ML algorithms for eHealth applications, based on data from [140].…”

A Machine-Learning-Based Analysis of the Relationships between Loneliness Metrics and Mobility Patterns for Elderly

“…In addition and considering the demography of the participants in our study (i.e., the elderly), it was necessary to select the two sensors as they offered the most convenient means for the elderly to report the current state of their mood at every point in time through the use of the push buttons. In order to analyze the data from these sensors, various machine-learning algorithms were also identified in our previous study [5]. The most used machine-learning algorithms in the literature for such studies are: the gradient boosting algorithms and the XGBoost which were used, for example, in the English Longitudinal Study of Ageing (ELSA) dataset to predict the loneliness [1].…”