Application of machine learning to predict transport modes from GPS, accelerometer, and heart rate data
Abstract Background There has been an increased focus on active transport, but the measurement of active transport is still difficult and error-prone. Sensor data have been used to predict active transport. While heart rate data have very rarely been considered before, this study used random forests...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2022-11-01
|
| Series: | International Journal of Health Geographics |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12942-022-00319-y |
| _version_ | 1797984041828352000 |
|---|---|
| author | Santosh Giri Ruben Brondeel Tarik El Aarbaoui Basile Chaix |
| author_facet | Santosh Giri Ruben Brondeel Tarik El Aarbaoui Basile Chaix |
| author_sort | Santosh Giri |
| collection | DOAJ |
| description | Abstract Background There has been an increased focus on active transport, but the measurement of active transport is still difficult and error-prone. Sensor data have been used to predict active transport. While heart rate data have very rarely been considered before, this study used random forests (RF) to predict transport modes using Global Positioning System (GPS), accelerometer, and heart rate data and paid attention to methodological issues related to the prediction strategy and post-processing. Methods The RECORD MultiSensor study collected GPS, accelerometer, and heart rate data over seven days from 126 participants living in the Ile-de-France region. RF models were built to predict transport modes for every minute (ground truth information on modes is from a GPS-based mobility survey), splitting observations between a Training dataset and a Test dataset at the participant level instead at the minute level. Moreover, several window sizes were tested for the post-processing moving average of the predicted transport mode. Results The minute-level prediction rate of being on trips vs. at a visited location was 90%. Final prediction rates of transport modes ranged from 65% for public transport to 95% for biking. Using minute-level observations from the same participants in the Training and Test sets (as RF spontaneously does) upwardly biases prediction rates. The inclusion of heart rate data improved prediction rates only for biking. A 3 to 5-min bandwidth moving average was optimum for a posteriori homogenization. Conclusion Heart rate only very slightly contributed to better predictions for specific transport modes. Moreover, our study shows that Training and Test sets must be carefully defined in RF models and that post-processing with carefully chosen moving average windows can improve predictions. |
| first_indexed | 2024-04-11T06:56:28Z |
| format | Article |
| id | doaj.art-318de4e87c9142dd8eee47a351cf0640 |
| institution | Directory Open Access Journal |
| issn | 1476-072X |
| language | English |
| last_indexed | 2024-04-11T06:56:28Z |
| publishDate | 2022-11-01 |
| publisher | BMC |
| record_format | Article |
| series | International Journal of Health Geographics |
| spelling | doaj.art-318de4e87c9142dd8eee47a351cf06402022-12-22T04:39:01ZengBMCInternational Journal of Health Geographics1476-072X2022-11-0121111410.1186/s12942-022-00319-yApplication of machine learning to predict transport modes from GPS, accelerometer, and heart rate dataSantosh Giri0Ruben Brondeel1Tarik El Aarbaoui2Basile Chaix3INSERM, Nemesis Research Team, Institut Pierre Louis d’Épidémiologie et de Santé Publique, Sorbonne UniversitéDepartment of Movement and Sport Sciences, Faculty of Medicine and Health Sciences, Ghent UniversityINSERM, Nemesis Research Team, Institut Pierre Louis d’Épidémiologie et de Santé Publique, Sorbonne UniversitéINSERM, Nemesis Research Team, Institut Pierre Louis d’Épidémiologie et de Santé Publique, Sorbonne UniversitéAbstract Background There has been an increased focus on active transport, but the measurement of active transport is still difficult and error-prone. Sensor data have been used to predict active transport. While heart rate data have very rarely been considered before, this study used random forests (RF) to predict transport modes using Global Positioning System (GPS), accelerometer, and heart rate data and paid attention to methodological issues related to the prediction strategy and post-processing. Methods The RECORD MultiSensor study collected GPS, accelerometer, and heart rate data over seven days from 126 participants living in the Ile-de-France region. RF models were built to predict transport modes for every minute (ground truth information on modes is from a GPS-based mobility survey), splitting observations between a Training dataset and a Test dataset at the participant level instead at the minute level. Moreover, several window sizes were tested for the post-processing moving average of the predicted transport mode. Results The minute-level prediction rate of being on trips vs. at a visited location was 90%. Final prediction rates of transport modes ranged from 65% for public transport to 95% for biking. Using minute-level observations from the same participants in the Training and Test sets (as RF spontaneously does) upwardly biases prediction rates. The inclusion of heart rate data improved prediction rates only for biking. A 3 to 5-min bandwidth moving average was optimum for a posteriori homogenization. Conclusion Heart rate only very slightly contributed to better predictions for specific transport modes. Moreover, our study shows that Training and Test sets must be carefully defined in RF models and that post-processing with carefully chosen moving average windows can improve predictions.https://doi.org/10.1186/s12942-022-00319-yTransport modePrediction modelsGlobal Positioning SystemAccelerometerHeart rateMachine Learning |
| spellingShingle | Santosh Giri Ruben Brondeel Tarik El Aarbaoui Basile Chaix Application of machine learning to predict transport modes from GPS, accelerometer, and heart rate data International Journal of Health Geographics Transport mode Prediction models Global Positioning System Accelerometer Heart rate Machine Learning |
| title | Application of machine learning to predict transport modes from GPS, accelerometer, and heart rate data |
| title_full | Application of machine learning to predict transport modes from GPS, accelerometer, and heart rate data |
| title_fullStr | Application of machine learning to predict transport modes from GPS, accelerometer, and heart rate data |
| title_full_unstemmed | Application of machine learning to predict transport modes from GPS, accelerometer, and heart rate data |
| title_short | Application of machine learning to predict transport modes from GPS, accelerometer, and heart rate data |
| title_sort | application of machine learning to predict transport modes from gps accelerometer and heart rate data |
| topic | Transport mode Prediction models Global Positioning System Accelerometer Heart rate Machine Learning |
| url | https://doi.org/10.1186/s12942-022-00319-y |
| work_keys_str_mv | AT santoshgiri applicationofmachinelearningtopredicttransportmodesfromgpsaccelerometerandheartratedata AT rubenbrondeel applicationofmachinelearningtopredicttransportmodesfromgpsaccelerometerandheartratedata AT tarikelaarbaoui applicationofmachinelearningtopredicttransportmodesfromgpsaccelerometerandheartratedata AT basilechaix applicationofmachinelearningtopredicttransportmodesfromgpsaccelerometerandheartratedata |