Automatic physical activity and in‐vehicle status classification based on GPS and accelerometer data: A hierarchical classification approach using machine learning techniques

2018 ◽  
Vol 22 (6) ◽  
pp. 1522-1549 ◽  
Author(s):  
Kangjae Lee ◽  
Mei‐Po Kwan
Keyword(s):  
Physical Activity ◽  
Machine Learning ◽  
Hierarchical Classification ◽  
Machine Learning Techniques ◽  
Accelerometer Data ◽  
Classification Approach ◽  
Learning Techniques ◽  
Status Classification ◽  
Vehicle Status

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 Hip and Wrist-Worn Accelerometer Data Analysis for Toddler Activities

2019 ◽  
Vol 16 (14) ◽  
pp. 2598 ◽  
Author(s):  
Soyang Kwon ◽  
Patricia Zavos ◽  
Katherine Nickele ◽  
Albert Sugianto ◽  
Mark V. Albert
Keyword(s):  
Physical Activity ◽  
Machine Learning ◽  
Sedentary Behavior ◽  
Vertical Axis ◽  
Free Play ◽  
Video Data ◽  
Machine Learning Techniques ◽  
Accelerometer Data ◽  
Signal Features ◽  
Learning Techniques

Although accelerometry data are widely utilized to estimate physical activity and sedentary behavior among children age 3 years or older, for toddlers age 1 and 2 year(s), accelerometry data recorded during such behaviors have been far less examined. In particular, toddler’s unique behaviors, such as riding in a stroller or being carried by an adult, have not yet been examined. The objective of this study was to describe accelerometry signal outputs recorded during participation in nine types of behaviors (i.e., running, walking, climbing up/down, crawling, riding a ride-on toy, standing, sitting, riding in a stroller/wagon, and being carried by an adult) among toddlers. Twenty-four toddlers aged 13 to 35 months (50% girls) performed various prescribed behaviors during free play in a commercial indoor playroom while wearing ActiGraph wGT3X-BT accelerometers on a hip and a wrist. Participants’ performances were video-recorded. Based on the video data, accelerometer data were annotated with behavior labels to examine accelerometry signal outputs while performing the nine types of behaviors. Accelerometer data collected during 664 behavior assessments from the 21 participants were used for analysis. Hip vertical axis counts for walking were low (median = 49 counts/5 s). They were significantly lower than those recorded while a toddler was “carried” by an adult (median = 144 counts/5 s; p < 0.01). While standing, sitting, and riding in a stroller, very low hip vertical axis counts were registered (median ≤ 5 counts/5 s). Although wrist vertical axis and vector magnitude counts for “carried” were not higher than those for walking, they were higher than the cut-points for sedentary behaviors. Using various accelerometry signal features, machine learning techniques showed 89% accuracy to differentiate the “carried” behavior from ambulatory movements such as running, walking, crawling, and climbing. In conclusion, hip vertical axis counts alone may be unable to capture walking as physical activity and “carried” as sedentary behavior among toddlers. Machine learning techniques that utilize additional accelerometry signal features could help to recognize behavior types, especially to differentiate being “carried” from ambulatory movements.

scholarly journals Physical activity prediction using fitness data: Challenges and issues

2021 ◽  
Vol 10 (1) ◽  
pp. 419-426
Author(s):  
Nur Zarna Elya Zakariya ◽  
Marshima Mohd Rosli
Keyword(s):  
Physical Activity ◽  
Machine Learning ◽  
Active Role ◽  
Physical Activities ◽  
Machine Learning Techniques ◽  
Behavior Model ◽  
Learning Techniques ◽  
Healthcare Technologies ◽  
Personal Context ◽  
Behavior Change Technique

In the new healthcare transformations, individuals are encourage to maintain healthy life based on their food diet and physical activity routine to avoid risk of serious disease. One of the recent healthcare technologies to support self health monitoring is wearable device that allow individual play active role on their own healthcare. However, there is still questions in terms of the accuracy of wearable data for recommending physical activity due to enormous fitness data generated by wearable devices. In this study, we conducted a literature review on machine learning techniques to predict suitable physical activities based on personal context and fitness data. We categorize and structure the research evidence that has been publish in the area of machine learning techniques for predicting physical activities using fitness data. We found 10 different models in Behavior Change Technique (BCT) and we selected two suitable models which are Fogg Behavior Model (FBM) and Trans-theoretical Behavior Model (TTM) for predicting physical activity using fitness data. We proposed a conceptual framework which consists of personal fitness data, combination of TTM and FBM to predict the suitable physical activity based on personal context. This study will provide new insights in software development of healthcare technologies to support personalization of individuals in managing their own health.

scholarly journals A Classification Approach for Predicting COVID-19 Patient's Survival Outcome with Machine Learning Techniques

2022 ◽  
Vol 22 (1) ◽  
pp. 1
Author(s):  
Abdulhameed Osi ◽  
Mannir Abdu ◽  
Hussaini Dikko ◽  
Usman Muhammad ◽  
Auwalu Ibrahim ◽  
...  
Keyword(s):  
Machine Learning ◽  
Survival Outcome ◽  
Machine Learning Techniques ◽  
Classification Approach ◽  
Learning Techniques

Machine learning applied to pack classification

2019 ◽  
Vol 61 (6) ◽  
pp. 601-620
Author(s):  
Weiqiang Hang ◽  
Timothy Banks
Keyword(s):  
Machine Learning ◽  
Market Research ◽  
Hierarchical Classification ◽  
Machine Learning Techniques ◽  
Support Vector ◽  
Sales Volume ◽  
Related Services ◽  
Learning Techniques ◽  
Common Task ◽  
Product Classification

Pack or product classification is a quite common task in market research, particularly for sales tracking audits and related services. Electronic data sources have led to increased volumes, both in the sales volume being tracked and also the number of packs (or stock keeping units). The increase in packs needing to be classified presents a problem, in that, it needs to be done accurately and quickly. Traditional solutions using people for the classifications can be costly, due to the large number of people required to process the classifications in a timely and accurate manner. Reducing the manual work is a priority for audit-based market research businesses, leading to interest in automation, such as through machine learning techniques. In this article, we apply such methods. These include support vector machine, decision tree, XGBoost, AdaBoost, random forest, and neural network–based methods that are trained on the textual descriptions of already classified packs. We also implement a hierarchical classification method to take advantage of the structure of classes of the products. Once the models are trained, they can be used on unclassified data. Where the methods are not confident in their classifications, humans can be asked to classify. The hope is that the methods can learn to classify accurately enough that the manual workloads are reduced to manageable levels. This article reviews various methods and then outlines tests using these methods on two datasets collected by Nielsen, showing good performance.

Sign in / Sign up

Export Citation Format

Share Document