Dynamic Prediction of Physical Exertion: Leveraging AI Models and Wearable Sensor Data During Cycling Exercise
<b>Background/Objectives</b>: This study aimed to explore machine learning approaches for predicting physical exertion using physiological signals collected from wearable devices. <b>Methods</b>: Both traditional machine learning and deep learning methods for classification a...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-12-01
|
| Series: | Diagnostics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-4418/15/1/52 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841549275273625600 |
|---|---|
| author | Aref Smiley Joseph Finkelstein |
| author_facet | Aref Smiley Joseph Finkelstein |
| author_sort | Aref Smiley |
| collection | DOAJ |
| description | <b>Background/Objectives</b>: This study aimed to explore machine learning approaches for predicting physical exertion using physiological signals collected from wearable devices. <b>Methods</b>: Both traditional machine learning and deep learning methods for classification and regression were assessed. The research involved 27 healthy participants engaged in controlled cycling exercises. Physiological data, including ECG, heart rate, oxygen saturation, and pedal speed (RPM), were collected during these sessions, which were divided into eight two-minute segments. Heart rate variability (HRV) was also calculated to serve as a predictive indicator. We employed two feature selection algorithms to identify the most relevant features for model training: Minimum Redundancy Maximum Relevance (MRMR) for both classification and regression, and Univariate Feature Ranking for Classification. A total of 34 traditional models were developed using MATLAB’s Classification Learner App, utilizing 20% of the data for testing. In addition, Long Short-Term Memory (LSTM) networks were trained on the top features selected by the MRMR and Univariate Feature Ranking algorithms to enhance model performance. Finally, the MRMR-selected features were used for regression to train the LSTM model for predicting continuous outcomes. <b>Results</b>: The LSTM model for regression demonstrated robust predictive capabilities, achieving a mean squared error (MSE) of 0.8493 and an R-squared value of 0.7757. The classification models also showed promising results, with the highest testing accuracy reaching 89.2% and an F1 score of 91.7%. <b>Conclusions</b>: These results underscore the effectiveness of combining feature selection algorithms with advanced machine learning (ML) and deep learning techniques for predicting physical exertion levels using wearable sensor data. |
| format | Article |
| id | doaj-art-6d169494ac0d4dc48e1430f6d39e5799 |
| institution | Kabale University |
| issn | 2075-4418 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Diagnostics |
| spelling | doaj-art-6d169494ac0d4dc48e1430f6d39e57992025-01-10T13:16:34ZengMDPI AGDiagnostics2075-44182024-12-011515210.3390/diagnostics15010052Dynamic Prediction of Physical Exertion: Leveraging AI Models and Wearable Sensor Data During Cycling ExerciseAref Smiley0Joseph Finkelstein1Department of Biomedical Informatics, University of Utah, Salt Lake City, UT 84108, USADepartment of Biomedical Informatics, University of Utah, Salt Lake City, UT 84108, USA<b>Background/Objectives</b>: This study aimed to explore machine learning approaches for predicting physical exertion using physiological signals collected from wearable devices. <b>Methods</b>: Both traditional machine learning and deep learning methods for classification and regression were assessed. The research involved 27 healthy participants engaged in controlled cycling exercises. Physiological data, including ECG, heart rate, oxygen saturation, and pedal speed (RPM), were collected during these sessions, which were divided into eight two-minute segments. Heart rate variability (HRV) was also calculated to serve as a predictive indicator. We employed two feature selection algorithms to identify the most relevant features for model training: Minimum Redundancy Maximum Relevance (MRMR) for both classification and regression, and Univariate Feature Ranking for Classification. A total of 34 traditional models were developed using MATLAB’s Classification Learner App, utilizing 20% of the data for testing. In addition, Long Short-Term Memory (LSTM) networks were trained on the top features selected by the MRMR and Univariate Feature Ranking algorithms to enhance model performance. Finally, the MRMR-selected features were used for regression to train the LSTM model for predicting continuous outcomes. <b>Results</b>: The LSTM model for regression demonstrated robust predictive capabilities, achieving a mean squared error (MSE) of 0.8493 and an R-squared value of 0.7757. The classification models also showed promising results, with the highest testing accuracy reaching 89.2% and an F1 score of 91.7%. <b>Conclusions</b>: These results underscore the effectiveness of combining feature selection algorithms with advanced machine learning (ML) and deep learning techniques for predicting physical exertion levels using wearable sensor data.https://www.mdpi.com/2075-4418/15/1/52heart rate variability (HRV)machine learningwearable sensorsphysical exertion prediction |
| spellingShingle | Aref Smiley Joseph Finkelstein Dynamic Prediction of Physical Exertion: Leveraging AI Models and Wearable Sensor Data During Cycling Exercise Diagnostics heart rate variability (HRV) machine learning wearable sensors physical exertion prediction |
| title | Dynamic Prediction of Physical Exertion: Leveraging AI Models and Wearable Sensor Data During Cycling Exercise |
| title_full | Dynamic Prediction of Physical Exertion: Leveraging AI Models and Wearable Sensor Data During Cycling Exercise |
| title_fullStr | Dynamic Prediction of Physical Exertion: Leveraging AI Models and Wearable Sensor Data During Cycling Exercise |
| title_full_unstemmed | Dynamic Prediction of Physical Exertion: Leveraging AI Models and Wearable Sensor Data During Cycling Exercise |
| title_short | Dynamic Prediction of Physical Exertion: Leveraging AI Models and Wearable Sensor Data During Cycling Exercise |
| title_sort | dynamic prediction of physical exertion leveraging ai models and wearable sensor data during cycling exercise |
| topic | heart rate variability (HRV) machine learning wearable sensors physical exertion prediction |
| url | https://www.mdpi.com/2075-4418/15/1/52 |
| work_keys_str_mv | AT arefsmiley dynamicpredictionofphysicalexertionleveragingaimodelsandwearablesensordataduringcyclingexercise AT josephfinkelstein dynamicpredictionofphysicalexertionleveragingaimodelsandwearablesensordataduringcyclingexercise |