scholarly journals PREDICTING THE TYPE OF PHYSICAL ACTIVITY FROM TRI-AXIAL SMARTPHONE ACCELEROMETER DATA

2021 ◽  
pp. 1-6
Author(s):  
Katarina Pavlović
Keyword(s):  
Physical Activity ◽  
Data Science ◽  
Human Movement ◽  
Supervised Machine Learning ◽  
Accelerometer Data ◽  
Digital Devices ◽  
Physical Mobility ◽  
Digital Phenotyping ◽  
Using Data ◽  
Randomized Search

Development of various statistical learning methods and their implementation in mobile device software enables moment-by-moment study of human social interactions, behavioral patterns, sleep, as well as their  physical mobility and gross motor activity. Recently, through the use of supervised Machine Learning, human activity recognition (HAR) has been found numerous applications in biomedical engineering especially in the field of digital phenotyping. Having this in mind, in this research in order to be able to quantify the human movement activity in situ, using data from portable digital devices,  we have developed code which uses Random Forest Classifier to predict the type of physical activity from tri-axial smartphone accelerometer data. The code has been written using Python programing language and Anaconda distribution of data-science packages. Raw accelerometer data was collected by using the Beiwe research platform, which is developed by the Onnela Lab at the Harvard T.H. Chan School of Public Health. Tuning has been performed by defining a grid of hyperparameter ranges, using Scikit-Learn’s Randomized Search CV method, randomly sampling from the grid and performing K-Fold CV with each combination of tested values. Obtained results will enable development a more robust models for predicting the type of physical activity with more subjects, usage of different hardwares, various test situations, and different environments.

scholarly journals The mobilize center: an NIH big data to knowledge center to advance human movement research and improve mobility

2015 ◽  
Vol 22 (6) ◽  
pp. 1120-1125 ◽  
Author(s):  
Joy P Ku ◽  
Jennifer L Hicks ◽  
Trevor Hastie ◽  
Jure Leskovec ◽  
Christopher Ré ◽  
...  
Keyword(s):  
Data Science ◽  
Human Mobility ◽  
Wearable Sensors ◽  
Human Movement ◽  
Science Research ◽  
Science Methods ◽  
Social Modeling ◽  
Mobility Data ◽  
Using Data ◽  
And Training

Abstract Regular physical activity helps prevent heart disease, stroke, diabetes, and other chronic diseases, yet a broad range of conditions impair mobility at great personal and societal cost. Vast amounts of data characterizing human movement are available from research labs, clinics, and millions of smartphones and wearable sensors, but integration and analysis of this large quantity of mobility data are extremely challenging. The authors have established the Mobilize Center ( http://mobilize.stanford.edu ) to harness these data to improve human mobility and help lay the foundation for using data science methods in biomedicine. The Center is organized around 4 data science research cores: biomechanical modeling, statistical learning, behavioral and social modeling, and integrative modeling. Important biomedical applications, such as osteoarthritis and weight management, will focus the development of new data science methods. By developing these new approaches, sharing data and validated software tools, and training thousands of researchers, the Mobilize Center will transform human movement research.

Analysis and visualization of COVID-19 discourse on Twitter using data science: a case study of the USA, the UK and India

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Haider Ilyas ◽  
Ahmed Anwar ◽  
Ussama Yaqub ◽  
Zamil Alzamil ◽  
Deniz Appelbaum
Keyword(s):  
Data Science ◽  
Supervised Machine Learning ◽  
Content Type ◽  
The Public ◽  
Public Sentiment ◽  
The People ◽  
The Usa ◽  
Using Data ◽  
The Uk

Purpose This paper aims to understand, examine and interpret the main concerns and emotions of the people regarding COVID-19 pandemic in the UK, the USA and India using Data Science measures. Design/methodology/approach This study implements unsupervised and supervised machine learning methods, i.e. topic modeling and sentiment analysis on Twitter data for extracting the topics of discussion and calculating public sentiment. Findings Governments and policymakers remained the focus of public discussion on Twitter during the first three months of the pandemic. Overall, public sentiment toward the pandemic remained neutral except for the USA. Originality/value This paper proposes a Data Science-based approach to better understand the public topics of concern during the COVID-19 pandemic.

Application of Raw Accelerometer Data and Machine-Learning Techniques to Characterize Human Movement Behavior: A Systematic Scoping Review

2020 ◽  
Vol 17 (3) ◽  
pp. 360-383 ◽  
Author(s):  
Anantha Narayanan ◽  
Farzanah Desai ◽  
Tom Stewart ◽  
Scott Duncan ◽  
Lisa Mackay
Keyword(s):  
Physical Activity ◽  
Machine Learning ◽  
Scoping Review ◽  
Machine Learning Algorithms ◽  
Supervised Machine Learning ◽  
Machine Learning Techniques ◽  
Movement Behavior ◽  
Support Vector ◽  
Accelerometer Data ◽  
Learning Techniques

Background: Application of machine learning for classifying human behavior is increasingly common as access to raw accelerometer data improves. The aims of this scoping review are (1) to examine if machine-learning techniques can accurately identify human activity behaviors from raw accelerometer data and (2) to summarize the practical implications of these machine-learning techniques for future work. Methods: Keyword searches were performed in Scopus, Web of Science, and EBSCO databases in 2018. Studies that applied supervised machine-learning techniques to raw accelerometer data and estimated components of physical activity were included. Information on study characteristics, machine-learning techniques, and key study findings were extracted from included studies. Results: Of the 53 studies included in the review, 75% were published in the last 5 years. Most studies predicted postures and activity type, rather than intensity, and were conducted in controlled environments using 1 or 2 devices. The most common models were support vector machine, random forest, and artificial neural network. Overall, classification accuracy ranged from 62% to 99.8%, although nearly 80% of studies achieved an overall accuracy above 85%. Conclusions: Machine-learning algorithms demonstrate good accuracy when predicting physical activity components; however, their application to free-living settings is currently uncertain.

scholarly journals Physical Activity and Health: Does the Pattern Matter?

2020 ◽  
Vol 4 (Supplement_1) ◽  
pp. 790-790
Author(s):  
Annemarie Koster ◽  
Sari Stenholm ◽  
Paul Gardiner
Keyword(s):  
Physical Activity ◽  
Sedentary Behavior ◽  
Latent Class ◽  
Activity Patterns ◽  
Health Study ◽  
Accelerometer Data ◽  
Aging Study ◽  
Health Related ◽  
The Usa ◽  
Using Data

Abstract It is well-known that physical activity is key in the prevention of many diseases and disability in old age. Much less is, however, known about the pattern of activity in relation to health. While there are differences in how people spread their activity and sedentary behavior over the day or over the week, we don’t know which activity pattern of most beneficial for health. This symposium focuses on patterns of physical activity and sedentary behavior and health in five different studies with accelerometry data in Europe and the USA. Dr. Rosenberg will show how sedentary behavior patterns are associated with various health outcomes in the Adult Changes in Thought (ACT) study. Using data from The Maastricht Study, Dr. Vandercappellen will present how weekly activity patterns, in particular comparing regularly actives to weekend warriors, are associated with arterial stiffness. Dr. Shiroma will show how patterns of physical activity and sedentary behavior, taking the volume, intensity, and frequency of sessions into account, are associated with mortality in the Women’s Health Study. Dr. Caserotti will present the association between physical activity fragmentation and physical function in the SITLESS Study. Dr. Stenholm will present data from the Finish Retirement and Aging Study, using latent class trajectory analyses to identify daily activity patterns and how these patterns are associated with health-related physical fitness. Taken together, this symposium will provide insight into different ways patterns of activity can be operationalized using accelerometer data and if the patterns of activity and sedentary behavior are associated with health.

The Effect of Sensor Placement and Number on Physical Activity Recognition and Energy Expenditure Estimation (Preprint)

2020 ◽  
Author(s):  
Anis Davoudi ◽  
Mamoun T. Mardini ◽  
Dave Nelson ◽  
Fahd Albinali ◽  
Sanjay Ranka ◽  
...  
Keyword(s):  
Physical Activity ◽  
Older Adults ◽  
Energy Expenditure ◽  
Activity Recognition ◽  
Recognition Accuracy ◽  
Body Position ◽  
Computer Work ◽  
Accelerometer Data ◽  
Upper Arm ◽  
Lifestyle Activities

BACKGROUND Research shows the feasibility of human activity recognition using Wearable accelerometer devices. Different studies have used varying number and placement for data collection using the sensors. OBJECTIVE To compare accuracy performance between multiple and variable placement of accelerometer devices in categorizing the type of physical activity and corresponding energy expenditure in older adults. METHODS Participants (n=93, 72.2±7.1 yrs) completed a total of 32 activities of daily life in a laboratory setting. Activities were classified as sedentary vs. non-sedentary, locomotion vs. non-locomotion, and lifestyle vs. non-lifestyle activities (e.g. leisure walk vs. computer work). A portable metabolic unit was worn during each activity to measure metabolic equivalents (METs). Accelerometers were placed on five different body positions: wrist, hip, ankle, upper arm, and thigh. Accelerometer data from each body position and combinations of positions were used in developing Random Forest models to assess activity category recognition accuracy and MET estimation. RESULTS Model performance for both MET estimation and activity category recognition strengthened with additional accelerometer devices. However, a single accelerometer on the ankle, upper arm, hip, thigh, or wrist had only a 0.03 to 0.09 MET increase in prediction error as compared to wearing all five devices. Balanced accuracy showed similar trends with slight decreases in balanced accuracy for detection of locomotion (0-0.01 METs), sedentary (0.13-0.05 METs) and lifestyle activities (0.08-0.04 METs) compared to all five placements. The accuracy of recognizing activity categories increased with additional placements (0.15-0.29). Notably, the hip was the best single body position for MET estimation and activity category recognition. CONCLUSIONS Additional accelerometer devices only slightly enhance activity recognition accuracy and MET estimation in older adults. However, given the extra burden of wearing additional devices, single accelerometers with appropriate placement appear to be sufficient for estimating energy expenditure and activity category recognition in older adults.

scholarly journals Physical Activity Patterns After Retirement: The REGARDS Study

2020 ◽  
Vol 4 (Supplement_1) ◽  
pp. 528-529
Author(s):  
Eric Shiroma ◽  
J David Rhodes ◽  
Aleena Bennet ◽  
Monika M Safford ◽  
Leslie MacDonald ◽  
...  
Keyword(s):  
Physical Activity ◽  
Light Intensity ◽  
Racial Differences ◽  
Activity Patterns ◽  
Accelerometer Data ◽  
Tv Viewing ◽  
Major Life Events ◽  
Major Life ◽  
Physical Activity Patterns ◽  
Regards Study

Abstract Major life events, such as retirement, may lead to dramatic shifts in physical activity (PA) patterns. However, there are limited empirical data quantifying the magnitude of these changes. Our aims were to objectively measure PA before and after retirement and to describe changes in participation in various types of PA. Participants were employed black and white men and women enrolled in REGARDS (REasons for Geographic and Racial Differences in Stroke), a national prospective cohort study (n=581, mean age 64 years, 25% black, 51% women). Participants met inclusion criteria if they retired between their first and second accelerometer wearing (2009-2013 and 2017-2018, respectively) and had valid accelerometer data (>4 days with >10 hours/day pre- and post-retirement). Accelerometer-based PA was categorized into average minutes per day spent in sedentary, light-intensity, and moderate-to-vigorous PA. Participants reported changes (less, same, more) in 12 types of PA. After retirement, participants decreased both sedentary time (by 36.3 minutes/day) and moderate-to-vigorous PA (by 5.6 minutes/day). Conversely, there was an increase in light-intensity PA (+18.1 minutes/day) after retirement. Participants reported changes in their participation level in various PA activities. For example, 41% reported an increased amount of TV viewing, 42% reported less walking, and 31% reported increased participation in volunteer activities. Findings indicate that retirement coincides with a change in the time spent in each intensity category and the time spent across a range of activity types. Further research is warranted to examine how these changes in physical activity patterns influence post-retirement health status.

Effect of Incentive Amount on U.S. Adolescents’ Participation in an Accelerometer Data Collection Component of a National Survey

Field Methods ◽  
2021 ◽  
pp. 1525822X2198984
Author(s):  
April Y. Oh ◽  
Andrew Caporaso ◽  
Terisa Davis ◽  
Laura A. Dwyer ◽  
Linda C. Nebeling ◽  
...  
Keyword(s):  
Physical Activity ◽  
Data Collection ◽  
National Survey ◽  
Monetary Incentive ◽  
Behavioral Research ◽  
Accelerometer Data ◽  
Web Based ◽  
Demographic Subgroups

Behavioral research increasingly uses accelerometers to provide objective estimates of physical activity. This study extends research on methods for collecting accelerometer data among youth by examining whether the amount of a monetary incentive affects enrollment and compliance in a mail-based accelerometer study of adolescents. We invited a subset of adolescents in a national web-based study to wear an accelerometer for seven days and return it by mail; participants received either $20 or $40 for participating. Enrollment did not significantly differ by incentive amount. However, adolescents receiving the $40 incentive had significantly higher compliance (accelerometer wear and return). This difference was largely consistent across demographic subgroups. Those in the $40 group also wore the accelerometer for more time than the $20 group on the first two days of the study. Compared to $20, a $40 incentive may promote youth completion of mail-based accelerometer studies.

scholarly journals Children and Adolescents Treated for Valvular Aortic Stenosis Have Different Physical Activity Patterns Compared to Healthy Controls: A Methodological Study in a National Cohort

2021 ◽  
Author(s):  
Pia Skovdahl ◽  
Cecilia Kjellberg Olofsson ◽  
Jan Sunnegårdh ◽  
Jonatan Fridolfsson ◽  
Mats Börjesson ◽  
...  
Keyword(s):  
Physical Activity ◽  
High Resolution ◽  
Aortic Stenosis ◽  
Children And Adolescents ◽  
Sedentary Time ◽  
Heart Defects ◽  
Sports Participation ◽  
Accelerometer Data ◽  
Healthy Controls ◽  
Valvular Aortic Stenosis

AbstractPrevious research in children and adolescents with congenital heart defects presents contradictory findings concerning their physical activity (PA) level, due to methodological limitations in the PA assessment. The aim of the present cross-sectional study was to compare PA in children and adolescents treated for valvular aortic stenosis with healthy controls using an improved accelerometer method. Seven-day accelerometer data were collected from the hip in a national Swedish sample of 46 patients 6–18 years old treated for valvular aortic stenosis and 44 healthy controls matched for age, gender, geography, and measurement period. Sports participation was self-reported. Accelerometer data were processed with the new improved Frequency Extended Method and with the traditional ActiGraph method for comparison. A high-resolution PA intensity spectrum was investigated as well as traditional crude PA intensity categories. Children treated for aortic stenosis had a pattern of less PA in the highest intensity spectra and had more sedentary time, while the adolescent patients tended to be less physically active in higher intensities overall and with less sedentary time, compared to the controls. These patterns were evident using the Frequency Extended Method with the detailed PA intensity spectrum, but not to the same degree using the ActiGraph method and traditional crude PA intensity categories. Patients reported less sports participation than their controls in both age-groups. Specific differences in PA patterns were revealed using the Frequency Extended Method with the high-resolution PA intensity spectrum in Swedish children and adolescents treated for valvular aortic stenosis.

scholarly journals Novel Analytic Approaches to Investigate Minute-Level Actigraphy and Associations With Physical Function

2020 ◽  
Vol 4 (Supplement_1) ◽  
pp. 194-195
Author(s):  
Kaiyuan Hua ◽  
Sheng Luo ◽  
Katherine Hall ◽  
Miriam Morey ◽  
Harvey Cohen
Keyword(s):  
Physical Activity ◽  
Data Analysis ◽  
Physical Function ◽  
Functional Data Analysis ◽  
Functional Data ◽  
Gait Speed ◽  
Community Dwelling ◽  
Accelerometer Data ◽  
Single Leg Stance ◽  
Minute Walk

Abstract Background. Functional decline in conjunction with low levels of physical activity has implications for health risks in older adults. Previous studies have examined the associations between accelerometry-derived activity and physical function, but most of these studies reduced these data into average means of total daily physical activity (e.g., daily step counts). A new method of analysis “functional data analysis” provides more in-depth capability using minute-level accelerometer data. Methods. A secondary analysis of community-dwelling adults ages 30 to 90+ residing in southwest region of North Carolina from the Physical Performance across the Lifespan (PALS) study. PALS assessments were completed in-person at baseline and one-week of accelerometry. Final analysis includes 669 observations at baseline with minute-level accelerometer data from 7:00 to 23:00, after removing non-wear time. A novel scalar-on-function regression analysis was used to explore the associations between baseline physical activity features (minute-by-minute vector magnitude generated from accelerometer) and baseline physical function (gait speed, single leg stance, chair stands, and 6-minute walk test) with control for baseline age, sex, race and body mass index. Results. The functional regressions were significant for specific times of day indicating increased physical activity associated with increased physical function around 8:00, 9:30 and 15:30-17:00 for rapid gait speed; 9:00-10:30 and 15:00-16:30 for normal gait speed; 9:00-10:30 for single leg stance; 9:30-11:30 and 15:00-18:00 for chair stands; 9:00-11:30 and 15:00-18:30 for 6-minute walk. Conclusion. This method of functional data analysis provides news insights into the relationship between minute-by-minute daily activity and health.

Sign in / Sign up

Export Citation Format

Share Document