scholarly journals An artificial neural network to estimate physical activity energy expenditure and identify physical activity type from an accelerometer

2009 ◽  
Vol 107 (4) ◽  
pp. 1300-1307 ◽  
Author(s):  
John Staudenmayer ◽  
David Pober ◽  
Scott Crouter ◽  
David Bassett ◽  
Patty Freedson
Keyword(s):  
Physical Activity ◽  
Oxygen Consumption ◽  
Confidence Interval ◽  
Mean Squared Error ◽  
Temporal Dynamics ◽  
Accelerometer Data ◽  
Ann Model ◽  
Activity Type ◽  
Metabolic Equivalents ◽  
Artificial Neural

The aim of this investigation was to develop and test two artificial neural networks (ANN) to apply to physical activity data collected with a commonly used uniaxial accelerometer. The first ANN model estimated physical activity metabolic equivalents (METs), and the second ANN identified activity type. Subjects ( n = 24 men and 24 women, mean age = 35 yr) completed a menu of activities that included sedentary, light, moderate, and vigorous intensities, and each activity was performed for 10 min. There were three different activity menus, and 20 participants completed each menu. Oxygen consumption (in ml·kg−1·min−1) was measured continuously, and the average of minutes 4–9 was used to represent the oxygen cost of each activity. To calculate METs, activity oxygen consumption was divided by 3.5 ml·kg−1·min−1 (1 MET). Accelerometer data were collected second by second using the Actigraph model 7164. For the analysis, we used the distribution of counts (10th, 25th, 50th, 75th, and 90th percentiles of a minute's second-by-second counts) and temporal dynamics of counts (lag, one autocorrelation) as the accelerometer feature inputs to the ANN. To examine model performance, we used the leave-one-out cross-validation technique. The ANN prediction of METs root-mean-squared error was 1.22 METs (confidence interval: 1.14–1.30). For the prediction of activity type, the ANN correctly classified activity type 88.8% of the time (confidence interval: 86.4–91.2%). Activity types were low-level activities, locomotion, vigorous sports, and household activities/other activities. This novel approach of applying ANNs for processing Actigraph accelerometer data is promising and shows that we can successfully estimate activity METs and identify activity type using ANN analytic procedures.

scholarly journals Evaluation of artificial neural network algorithms for predicting METs and activity type from accelerometer data: validation on an independent sample

2011 ◽  
Vol 111 (6) ◽  
pp. 1804-1812 ◽  
Author(s):  
Patty S. Freedson ◽  
Kate Lyden ◽  
Sarah Kozey-Keadle ◽  
John Staudenmayer
Keyword(s):  
Machine Learning ◽  
Training Data ◽  
Accelerometer Data ◽  
Activity Type ◽  
Data Set ◽  
University Of Tennessee ◽  
Metabolic Equivalents ◽  
Artificial Neural ◽  
University Of Massachusetts ◽  
Mean Square Errors

Previous work from our laboratory provided a “proof of concept” for use of artificial neural networks (nnets) to estimate metabolic equivalents (METs) and identify activity type from accelerometer data (Staudenmayer J, Pober D, Crouter S, Bassett D, Freedson P, J Appl Physiol 107: 1330–1307, 2009). The purpose of this study was to develop new nnets based on a larger, more diverse, training data set and apply these nnet prediction models to an independent sample to evaluate the robustness and flexibility of this machine-learning modeling technique. The nnet training data set (University of Massachusetts) included 277 participants who each completed 11 activities. The independent validation sample ( n = 65) (University of Tennessee) completed one of three activity routines. Criterion measures were 1) measured METs assessed using open-circuit indirect calorimetry; and 2) observed activity to identify activity type. The nnet input variables included five accelerometer count distribution features and the lag-1 autocorrelation. The bias and root mean square errors for the nnet MET trained on University of Massachusetts and applied to University of Tennessee were +0.32 and 1.90 METs, respectively. Seventy-seven percent of the activities were correctly classified as sedentary/light, moderate, or vigorous intensity. For activity type, household and locomotion activities were correctly classified by the nnet activity type 98.1 and 89.5% of the time, respectively, and sport was correctly classified 23.7% of the time. Use of this machine-learning technique operates reasonably well when applied to an independent sample. We propose the creation of an open-access activity dictionary, including accelerometer data from a broad array of activities, leading to further improvements in prediction accuracy for METs, activity intensity, and activity type.

scholarly journals Behavioural interventions to promote physical activity in a multiethnic population at high risk of diabetes: PROPELS three-arm RCT

2021 ◽  
Vol 25 (77) ◽  
pp. 1-190
Author(s):  
Kamlesh Khunti ◽  
Simon Griffin ◽  
Alan Brennan ◽  
Helen Dallosso ◽  
Melanie Davies ◽  
...  
Keyword(s):  
Physical Activity ◽  
Type 2 Diabetes ◽  
Confidence Interval ◽  
Primary Outcome ◽  
Health Technology ◽  
Accelerometer Data ◽  
Ambulatory Activity ◽  
Steps Per Day

Background Type 2 diabetes is a leading cause of mortality globally and accounts for significant health resource expenditure. Increased physical activity can reduce the risk of diabetes. However, the longer-term clinical effectiveness and cost-effectiveness of physical activity interventions in those at high risk of type 2 diabetes is unknown. Objectives To investigate whether or not Walking Away from Diabetes (Walking Away) – a low-resource, 3-hour group-based behavioural intervention designed to promote physical activity through pedometer use in those with prediabetes – leads to sustained increases in physical activity when delivered with and without an integrated mobile health intervention compared with control. Design Three-arm, parallel-group, pragmatic, superiority randomised controlled trial with follow-up conducted at 12 and 48 months. Setting Primary care and the community. Participants Adults whose primary care record included a prediabetic blood glucose measurement recorded within the past 5 years [HbA1c ≥ 42 mmol/mol (6.0%), < 48 mmol/mol (6.5%) mmol/mol; fasting glucose ≥ 5.5 mmol/l, < 7.0 mmol/l; or 2-hour post-challenge glucose ≥ 7.8 mmol/l, < 11.1 mmol/l] were recruited between December 2013 and February 2015. Data collection was completed in July 2019. Interventions Participants were randomised (1 : 1 : 1) using a web-based tool to (1) control (information leaflet), (2) Walking Away with annual group-based support or (3) Walking Away Plus (comprising Walking Away, annual group-based support and a mobile health intervention that provided automated, individually tailored text messages to prompt pedometer use and goal-setting and provide feedback, in addition to biannual telephone calls). Participants and data collectors were not blinded; however, the staff who processed the accelerometer data were blinded to allocation. Main outcome measures The primary outcome was accelerometer-measured ambulatory activity (steps per day) at 48 months. Other objective and self-reported measures of physical activity were also assessed. Results A total of 1366 individuals were randomised (median age 61 years, median body mass index 28.4 kg/m2, median ambulatory activity 6638 steps per day, women 49%, black and minority ethnicity 28%). Accelerometer data were available for 1017 (74%) and 993 (73%) individuals at 12 and 48 months, respectively. The primary outcome assessment at 48 months found no differences in ambulatory activity compared with control in either group (Walking Away Plus: 121 steps per day, 97.5% confidence interval –290 to 532 steps per day; Walking Away: 91 steps per day, 97.5% confidence interval –282 to 463). This was consistent across ethnic groups. At the intermediate 12-month assessment, the Walking Away Plus group had increased their ambulatory activity by 547 (97.5% confidence interval 211 to 882) steps per day compared with control and were 1.61 (97.5% confidence interval 1.05 to 2.45) times more likely to achieve 150 minutes per week of objectively assessed unbouted moderate to vigorous physical activity. In the Walking Away group, there were no differences compared with control at 12 months. Secondary anthropometric, biomechanical and mental health outcomes were unaltered in either intervention study arm compared with control at 12 or 48 months, with the exception of small, but sustained, reductions in body weight in the Walking Away study arm (≈ 1 kg) at the 12- and 48-month follow-ups. Lifetime cost-effectiveness modelling suggested that usual care had the highest probability of being cost-effective at a threshold of £20,000 per quality-adjusted life-year. Of 50 serious adverse events, only one (myocardial infarction) was deemed possibly related to the intervention and led to the withdrawal of the participant from the study. Limitations Loss to follow-up, although the results were unaltered when missing data were replaced using multiple imputation. Conclusions Combining a physical activity intervention with text messaging and telephone support resulted in modest, but clinically meaningful, changes in physical activity at 12 months, but the changes were not sustained at 48 months. Future work Future research is needed to investigate which intervention types, components and features can help to maintain physical activity behaviour change over the longer term. Trial registration Current Controlled Trials ISRCTN83465245. Funding This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 25, No. 77. See the NIHR Journals Library website for further project information.

scholarly journals Predictive Analysis of Employee Loyalty: A Comparative Study Using Logistic Regression Model and Artificial Neural Network

2019 ◽  
Vol 25 (3) ◽  
pp. 325-335
Author(s):  
Maria Zefanya Sampe ◽  
Eko Ariawan ◽  
I Wayan Ariawan
Keyword(s):  
Logistic Regression ◽  
Employee Turnover ◽  
Mean Squared Error ◽  
Working Hours ◽  
Retention Rates ◽  
Ann Model ◽  
Employee Loyalty ◽  
High Turnover ◽  
Kolmogorov Smirnov ◽  
Artificial Neural

Employee turnover is a common issue in any company. A high turnover phenomenon becomes a big problem that will certainly affect the performance of the company. Therefore, measuring employee turnover can be helpful to employers to improve employee retention rates and give them a head start on turnover. A study to analyze for employee loyalty has been carried out by using Logistic Regression (LR) and Artificial Neural Networks (ANN) model. Response variables such as satisfaction level, number of projects, average monthly working hours, employment period, working accident, promotion in the last 5 years, department, and salary level are used to model the employee turnover. Parameters such as accuracy, precision, sensitivity, Kolmogorov-Smirnov statistic, and Mean Squared Error (MSE) are used to compare both models.

scholarly journals Towards a Portable Model to Discriminate Activity Clusters from Accelerometer Data

Sensors ◽  
2019 ◽  
Vol 19 (20) ◽  
pp. 4504 ◽  
Author(s):  
Petra Jones ◽  
Evgeny M. Mirkes ◽  
Tom Yates ◽  
Charlotte L. Edwardson ◽  
Mike Catt ◽  
...  
Keyword(s):  
Physical Activity ◽  
Cross Validation ◽  
Accelerometer Data ◽  
Activity Type ◽  
Free Living ◽  
Independent Dataset ◽  
Clustering Model ◽  
Activity Types ◽  
Unsupervised Approach ◽  
Viable Approach

Few methods for classifying physical activity from accelerometer data have been tested using an independent dataset for cross-validation, and even fewer using multiple independent datasets. The aim of this study was to evaluate whether unsupervised machine learning was a viable approach for the development of a reusable clustering model that was generalisable to independent datasets. We used two labelled adult laboratory datasets to generate a k-means clustering model. To assess its generalised application, we applied the stored clustering model to three independent labelled datasets: two laboratory and one free-living. Based on the development labelled data, the ten clusters were collapsed into four activity categories: sedentary, standing/mixed/slow ambulatory, brisk ambulatory, and running. The percentages of each activity type contained in these categories were 89%, 83%, 78%, and 96%, respectively. In the laboratory independent datasets, the consistency of activity types within the clusters dropped, but remained above 70% for the sedentary clusters, and 85% for the running and ambulatory clusters. Acceleration features were similar within each cluster across samples. The clusters created reflected activity types known to be associated with health and were reasonably robust when applied to diverse independent datasets. This suggests that an unsupervised approach is potentially useful for analysing free-living accelerometer data.

Surface Roughness Prediction in Turning of Free Machining Steel 1215 by Artificial Neural Network

2011 ◽  
Vol 188 ◽  
pp. 535-541
Author(s):  
Xiao Jiang Cai ◽  
Z.Q. Liu ◽  
Q.C. Wang ◽  
Shu Han ◽  
Qing Long An ◽  
...  
Keyword(s):  
Neural Network ◽  
Surface Roughness ◽  
Artificial Neural Network ◽  
Prediction Model ◽  
Mean Squared Error ◽  
Cutting Speed ◽  
Depth Of Cut ◽  
Ann Model ◽  
Roughness Prediction ◽  
Artificial Neural

Surface roughness is a significant aspect of the surface integrity concept. It is efficient to predict the surface roughness in advance by a prediction model. In this study, artificial neural network is used to model the surface roughness in turning of free machining steel 1215. The inputs considered in the prediction ANN model were cutting speed, feed rate and depth of cut, and the output was Ra. Several feed-forward neural networks with different architectures were compared in terms of prediction accuracy, and then the best prediction model, a 3-4-1-1 ANN was capable of predicting Ra with a mean squared error 5.46%, was presented.

scholarly journals Artificial Neural Network Model for Managing and Forecasting Water Reservoir Discharge (Hemren Reservoir as A Case Study)

2014 ◽  
Vol 7 (4) ◽  
pp. 132-143
Author(s):  
ABBAS M. ABD ◽  
SAAD SH. SAMMEN
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Correlation Coefficient ◽  
Mean Squared Error ◽  
Water Reservoir ◽  
Prediction Method ◽  
Training Data ◽  
Ann Model ◽  
Data Set ◽  
Artificial Neural

The prediction of different hydrological phenomenon (or system) plays an increasing role in the management of water resources. As engineers; it is required to predict the component of natural reservoirs’ inflow for numerous purposes. Resulting prediction techniques vary with the potential purpose, characteristics, and documented data. The best prediction method is of interest of experts to overcome the uncertainty, because the most hydrological parameters are subjected to the uncertainty. Artificial Neural Network (ANN) approach has adopted in this paper to predict Hemren reservoir inflow. Available data including monthly discharge supplied from DerbendiKhan reservoir and rain fall intensity falling on the intermediate catchment area between Hemren-DerbendiKhan dams were used.A Back Propagation (LMBP) algorithm (Levenberg-Marquardt) has been utilized to construct the ANN models. For the developed ANN model, different networks with different numbers of neurons and layers were evaluated. A total of 24 years of historical data for interval from 1980 to 2004 were used to train and test the networks. The optimum ANN network with 3 inputs, 40 neurons in both two hidden layers and one output was selected. Mean Squared Error (MSE) and the Correlation Coefficient (CC) were employed to evaluate the accuracy of the proposed model. The network was trained and converged at MSE = 0.027 by using training data subjected to early stopping approach. The network could forecast the testing data set with the accuracy of MSE = 0.031. Training and testing process showed the correlation coefficient of 0.97 and 0.77 respectively and this is refer to a high precision of that prediction technique.

scholarly journals Ability of RT3 Accelerometer Cut Points to Detect Physical Activity Intensity in Ambulatory Children With Cerebral Palsy

2014 ◽  
Vol 31 (4) ◽  
pp. 310-324 ◽  
Author(s):  
Jennifer Ryan ◽  
Michael Walsh ◽  
John Gormley
Keyword(s):  
Physical Activity ◽  
Cerebral Palsy ◽  
Vigorous Physical Activity ◽  
Accelerometer Data ◽  
Physical Activity Intensity ◽  
Cut Points ◽  
Cut Point ◽  
Activity Intensity ◽  
Metabolic Equivalents ◽  
Children With Cerebral Palsy

This study investigated the ability of published cut points for the RT3 accelerometer to differentiate between levels of physical activity intensity in children with cerebral palsy (CP). Oxygen consumption (metabolic equivalents; METs) and RT3 data (counts/min) were measured during rest and 5 walking trials. METs and corresponding counts/min were classified as sedentary, light physical activity (LPA), and moderate to vigorous physical activity (MVPA) according to MET thresholds. Counts were also classified according to published cut points. A published cut point exhibited an excellent ability to classify sedentary activity (sensitivity = 89.5%, specificity = 100.0%). Classification accuracy decreased when published cut points were used to classify LPA (sensitivity = 88.9%, specificity = 79.6%) and MVPA (sensitivity = 70%, specificity = 95–97%). Derivation of a new cut point improved classification of both LPA and MVPA. Applying published cut points to RT3 accelerometer data collected in children with CP may result in misclassification of LPA and MVPA.

scholarly journals Distributed lag and spline modeling for predicting energy expenditure from accelerometry in youth

2010 ◽  
Vol 108 (2) ◽  
pp. 314-327 ◽  
Author(s):  
Leena Choi ◽  
Kong Y. Chen ◽  
Sari A. Acra ◽  
Maciej S. Buchowski
Keyword(s):  
Physical Activity ◽  
Energy Expenditure ◽  
Prediction Models ◽  
Prediction Errors ◽  
Accelerometer Data ◽  
Modeling Approach ◽  
Distributed Lag ◽  
Metabolic Equivalents ◽  
Accuracy And Precision ◽  
Better Than

Movement sensing using accelerometers is commonly used for the measurement of physical activity (PA) and estimating energy expenditure (EE) under free-living conditions. The major limitation of this approach is lack of accuracy and precision in estimating EE, especially in low-intensity activities. Thus the objective of this study was to investigate benefits of a distributed lag spline (DLS) modeling approach for the prediction of total daily EE (TEE) and EE in sedentary (1.0–1.5 metabolic equivalents; MET), light (1.5–3.0 MET), and moderate/vigorous (≥3.0 MET) intensity activities in 10- to 17-year-old youth ( n = 76). We also explored feasibility of the DLS modeling approach to predict physical activity EE (PAEE) and METs. Movement was measured by Actigraph accelerometers placed on the hip, wrist, and ankle. With whole-room indirect calorimeter as the reference standard, prediction models ( Hip, Wrist, Ankle, Hip+ Wrist, Hip+ Wrist+ Ankle) for TEE, PAEE, and MET were developed and validated using the fivefold cross-validation method. The TEE predictions by these DLS models were not significantly different from the room calorimeter measurements (all P > 0.05). The Hip+ Wrist+ Ankle predicted TEE better than other models and reduced prediction errors in moderate/vigorous PA for TEE, MET, and PAEE (all P < 0.001). The Hip+ Wrist reduced prediction errors for the PAEE and MET at sedentary PA ( P = 0.020 and 0.021) compared with the Hip. Models that included Wrist correctly classified time spent at light PA better than other models. The means and standard deviations of the prediction errors for the Hip+ Wrist+ Ankle and Hip were 0.4 ± 144.0 and 1.5 ± 164.7 kcal for the TEE, 0.0 ± 84.2 and 1.3 ± 104.7 kcal for the PAEE, and −1.1 ± 97.6 and −0.1 ± 108.6 MET min for the MET models. We conclude that the DLS approach for accelerometer data improves detailed EE prediction in youth.

Artificial neural networks modelling based on visual analysis of coated cross laminated timber (CLT) to predict color change during outdoor exposure

Holzforschung ◽  
2020 ◽  
Vol 0 (0) ◽  
Author(s):  
Gabrielly S. Bobadilha ◽  
C. Elizabeth Stokes ◽  
Dercilio Junior Verly Lopes
Keyword(s):  
Visual Analysis ◽  
Mean Squared Error ◽  
Color Change ◽  
Visual Assessment ◽  
Coefficient Of Determination ◽  
Ann Model ◽  
Appearance Time ◽  
Cross Laminated Timber ◽  
Unseen Data ◽  
Artificial Neural

AbstractIn this study, an artificial neural network (ANN) model was designed to predict color change based on visual assessment of coated cross laminated timber (CLT) exposed outdoors. Coatings and stains were investigated based on ASTM protocols to assess wood surface visual rating, against checking, flaking, erosion, and mildew growth in the State of Mississippi (USA) during one year (2019–2020). It was hypothesized that accurate ratings would promote precise color prediction by the ANN model. Visual assessment inputs were used to develop the model for predicting total color change (ΔE). The training and validation splits of the network were based on a 10-fold cross-validation technique, and the ANN model performance was assessed on the validation set using mean squared error (MSE), mean average precision (MAE), and coefficient of determination (R2) after permutation feature importance analysis (PFI). Results indicated that coating was the most important feature in color change model. Erosion, checking and flaking achieved similar importance with an approximate difference of 6%. The ANN model was able to effectively predict color change values based on visual ratings with overall accuracy of 95% on truly unseen data. These findings revealed that coating properties, visual appearance, time of exposure, are associated with discoloration. Accurate visual assessment and a well-trained ANN can successfully provide the desired values of ΔE with a smaller number of complex test procedures.

scholarly journals Decision support system for tool condition monitoring in milling process using artificial neural network

2021 ◽  
Author(s):  
.Mohanraj T ◽  
◽  
Tamilvanan A. ◽  
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Condition Monitoring ◽  
Mean Squared Error ◽  
Flank Wear ◽  
Tool Condition Monitoring ◽  
Ann Model ◽  
Vibration Signals ◽  
Tool Condition ◽  
Artificial Neural

This work discusses the development of tool condition monitoring system (TCMs) during milling of AISI stainless steel 304 using sound pressure and vibration signals. Response Surface Methodology (RSM) was used to design the experiments. The various milling parameters and vegetable-based cutting fluids (VBCFs) were optimized to reduce the surface roughness and flank wear. The experimental results reveal the direct relationship between the flank wear and sound and vibration signals. The various statistical parameters were extracted from the measured signals and given as input data to train the artificial neural network (ANN). From the developed ANN model, the flank wear was predicted with the mean squared error (MSE) of 0.0656 mm.

Sign in / Sign up

Export Citation Format

Share Document