Achieving Accelerometer Wrist and Orientation Invariance in Physical Activity Classification via Domain Adaption

2019 ◽  
Vol 2 (4) ◽  
pp. 256-262
Author(s):  
Joshua Twaites ◽  
Richard Everson ◽  
Joss Langford ◽  
Melvyn Hillsdon
Keyword(s):  
Physical Activity ◽  
Classification Models ◽  
Activity Classification ◽  
Domain Adaption ◽  
Sensor Orientation ◽  
Testing Data ◽  
Valuable Method ◽  
Using Data ◽  
Orientation Invariance

Purpose: Physical activity classifiers are typically trained on data obtained from sensors at a set orientation. Changes in this orientation (such as being on a different wrist) result in performance degradation. This work investigates a method to obtain sensor location and orientation invariance for classification of wrist-mounted accelerometry via a technique known as domain adaption. Methods: Data was gathered from 16 participants who wore accelerometers on both wrists. Physical activity classification models were created using data from each wrist and then used to predict activities when using data from the opposing wrist. Using subspace alignment domain adaption, this procedure was then repeated to align the training and testing data before the classification stage. Results: Prediction of activity when using data where the wearer’s wrist was incorrectly specified resulted in a significant (p = .01) decrease in performance of 12%. When using domain adaption this drop in performance became negligible (M difference < 1%, p = .73). Conclusion: Domain adaption is a valuable method for achieving accurate physical activity classification independent of sensor orientation in wrist-worn accelerometry.

scholarly journals Can machine learning augment clinician adjudication of events in cardiovascular trials? A case study of major adverse cardiovascular events (MACE) across CVRM trials

2021 ◽  
Vol 42 (Supplement_1) ◽  
Author(s):  
H Lea ◽  
E Hutchinson ◽  
A Meeson ◽  
S Nampally ◽  
G Dennis ◽  
...  
Keyword(s):  
Machine Learning ◽  
Clinical Trials ◽  
Ischemic Stroke ◽  
Cardiovascular Events ◽  
Learning Algorithms ◽  
Model Performance ◽  
Machine Learning Algorithms ◽  
Classification Models ◽  
Using Data

Abstract Background and introduction Accurate identification of clinical outcome events is critical to obtaining reliable results in cardiovascular outcomes trials (CVOTs). Current processes for event adjudication are expensive and hampered by delays. As part of a larger project to more reliably identify outcomes, we evaluated the use of machine learning to automate event adjudication using data from the SOCRATES trial (NCT01994720), a large randomized trial comparing ticagrelor and aspirin in reducing risk of major cardiovascular events after acute ischemic stroke or transient ischemic attack (TIA). Purpose We studied whether machine learning algorithms could replicate the outcome of the expert adjudication process for clinical events of ischemic stroke and TIA. Could classification models be trained on historical CVOT data and demonstrate performance comparable to human adjudicators? Methods Using data from the SOCRATES trial, multiple machine learning algorithms were tested using grid search and cross validation. Models tested included Support Vector Machines, Random Forest and XGBoost. Performance was assessed on a validation subset of the adjudication data not used for training or testing in model development. Metrics used to evaluate model performance were Receiver Operating Characteristic (ROC), Matthews Correlation Coefficient, Precision and Recall. The contribution of features, attributes of data used by the algorithm as it is trained to classify an event, that contributed to a classification were examined using both Mutual Information and Recursive Feature Elimination. Results Classification models were trained on historical CVOT data using adjudicator consensus decision as the ground truth. Best performance was observed on models trained to classify ischemic stroke (ROC 0.95) and TIA (ROC 0.97). Top ranked features that contributed to classification of Ischemic Stroke or TIA corresponded to site investigator decision or variables used to define the event in the trial charter, such as duration of symptoms. Model performance was comparable across the different machine learning algorithms tested with XGBoost demonstrating the best ROC on the validation set for correctly classifying both stroke and TIA. Conclusions Our results indicate that machine learning may augment or even replace clinician adjudication in clinical trials, with potential to gain efficiencies, speed up clinical development, and retain reliability. Our current models demonstrate good performance at binary classification of ischemic stroke and TIA within a single CVOT with high consistency and accuracy between automated and clinician adjudication. Further work will focus on harmonizing features between multiple historical clinical trials and training models to classify several different endpoint events across trials. Our aim is to utilize these clinical trial datasets to optimize the delivery of CVOTs in further cardiovascular drug development. FUNDunding Acknowledgement Type of funding sources: Private company. Main funding source(s): AstraZenca Plc

scholarly journals Teknik Bagging Dan Boosting Pada Algoritma CART Untuk Klasifikasi Masa Studi Mahasiswa

2019 ◽  
Vol 5 (1) ◽  
pp. 21-30
Author(s):  
Ahmad Rusadi Arrahimi ◽  
Muhammad Khairi Ihsan ◽  
Dwi Kartini ◽  
Mohammad Reza Faisal ◽  
Fatma Indriani
Keyword(s):  
Data Mining ◽  
Information Systems ◽  
Undergraduate Students ◽  
New Knowledge ◽  
Testing Data ◽  
Academic Information ◽  
Using Data ◽  
Boosting Technique ◽  
Cart Algorithm

Undergraduate Students data in academic information systems always increases every year. Data collected can be processed using data mining to gain new knowledge. The author tries to mine undergraduate students data to classify the study period on time or not on time. The data is analyzed using CART with bagging techniqu, and CART with boosting technique. The classification results using 49 testing data, in the CART algorithm with bagging techniques 13 data (26.531%) entered into the classification on time and 36 data (73.469%) entered into the classification not on time. In the CART algorithm with boosting technique 16 data (32,653%) entered into the classification on time and 33 data (67,347%) entered into the classification not on time. The accuracy value of the classification of study period of undergraduate students using the CART algorithm is 79.592%, the CART algorithm with bagging technique is 81.633%, and the CART algorithm with boosting technique is 87.755%. In this study, the CART algorithm with boosting technique has the best accuracy value.

Effectiveness of Supervised Classification Models for Hate Speech on Twitter

2020 ◽  
Author(s):  
Kunal Srivastava ◽  
Ryan Tabrizi ◽  
Ayaan Rahim ◽  
Lauryn Nakamitsu
Keyword(s):  
Supervised Classification ◽  
Hate Speech ◽  
The Internet ◽  
Classification Models ◽  
Personal Attack ◽  
Political Beliefs ◽  
The Many ◽  
Primary Medium ◽  
Offensive Speech

<div> <div> <div> <p>Abstract </p> <p>The ceaseless connectivity imposed by the internet has made many vulnerable to offensive comments, be it their physical appearance, political beliefs, or religion. Some define hate speech as any kind of personal attack on one’s identity or beliefs. Of the many sites that grant the ability to spread such offensive speech, Twitter has arguably become the primary medium for individuals and groups to spread these hurtful comments. Such comments typically fail to be detected by Twitter’s anti-hate system and can linger online for hours before finally being taken down. Through sentiment analysis, this algorithm is able to distinguish hate speech effectively through the classification of sentiment. </p> </div> </div> </div>

scholarly journals UAV Image Multi-Labeling with Data-Efficient Transformers

2021 ◽  
Vol 11 (9) ◽  
pp. 3974
Author(s):  
Laila Bashmal ◽  
Yakoub Bazi ◽  
Mohamad Mahmoud Al Rahhal ◽  
Haikel Alhichri ◽  
Naif Al Ajlan
Keyword(s):  
Data Augmentation ◽  
Feature Representation ◽  
Aerial Image ◽  
Remote Sensing Images ◽  
Training Set ◽  
Proposed Model ◽  
Class Labels ◽  
Using Data ◽  
Uav Image

In this paper, we present an approach for the multi-label classification of remote sensing images based on data-efficient transformers. During the training phase, we generated a second view for each image from the training set using data augmentation. Then, both the image and its augmented version were reshaped into a sequence of flattened patches and then fed to the transformer encoder. The latter extracts a compact feature representation from each image with the help of a self-attention mechanism, which can handle the global dependencies between different regions of the high-resolution aerial image. On the top of the encoder, we mounted two classifiers, a token and a distiller classifier. During training, we minimized a global loss consisting of two terms, each corresponding to one of the two classifiers. In the test phase, we considered the average of the two classifiers as the final class labels. Experiments on two datasets acquired over the cities of Trento and Civezzano with a ground resolution of two-centimeter demonstrated the effectiveness of the proposed model.

scholarly journals Lockdown, slow down: impact of the COVID-19 pandemic on physical activity—an observational study

Open Heart ◽  
2021 ◽  
Vol 8 (1) ◽  
pp. e001600
Author(s):  
Joanne Kathryn Taylor ◽  
Haarith Ndiaye ◽  
Matthew Daniels ◽  
Fozia Ahmed
Keyword(s):  
Physical Activity ◽  
New York ◽  
Heart Association ◽  
Evaluation Study ◽  
Activity Data ◽  
Cardiac Devices ◽  
Significant Difference ◽  
New York Heart Association ◽  
Using Data ◽  
The Uk

AimsIn response to the COVID-19 pandemic, the UK was placed under strict lockdown measures on 23 March 2020. The aim of this study was to quantify the effects on physical activity (PA) levels using data from the prospective Triage-HF Plus Evaluation study.MethodsThis study represents a cohort of adult patients with implanted cardiac devices capable of measuring activity by embedded accelerometery via a remote monitoring platform. Activity data were available for the 4 weeks pre-implementation and post implementation of ‘stay at home’ lockdown measures in the form of ‘minutes active per day’ (min/day).ResultsData were analysed for 311 patients (77.2% men, mean age 68.8, frailty 55.9%. 92.2% established heart failure (HF) diagnosis, of these 51.2% New York Heart Association II), with comorbidities representative of a real-world cohort.Post-lockdown, a significant reduction in median PA equating to 20.8 active min/day was seen. The reduction was uniform with a slightly more pronounced drop in PA for women, but no statistically significant difference with respect to age, body mass index, frailty or device type. Activity dropped in the immediate 2-week period post-lockdown, but steadily returned thereafter. Median activity week 4 weeks post-lockdown remained significantly lower than 4 weeks pre-lockdown (p≤0.001).ConclusionsIn a population of predominantly HF patients with cardiac devices, activity reduced by approximately 20 min active per day in the immediate aftermath of strict COVID-19 lockdown measures.Trial registration numberNCT04177199.

scholarly journals Multivariate Analysis for the Classification of Chocolate According to its Percentage of Cocoa by Using Terahertz Time-Domain Spectroscopy (THz-TDS)

Proceedings ◽  
2020 ◽  
Vol 70 (1) ◽  
pp. 109
Author(s):  
Jimy Oblitas ◽  
Jorge Ruiz
Keyword(s):  
Machine Learning ◽  
Time Domain ◽  
Electromagnetic Pulse ◽  
Machine Learning Algorithms ◽  
Classification Models ◽  
Terahertz Time Domain Spectroscopy ◽  
Time Domain Spectroscopy ◽  
Svm Algorithm ◽  
Classification Of Images

Terahertz time-domain spectroscopy is a useful technique for determining some physical characteristics of materials, and is based on selective frequency absorption of a broad-spectrum electromagnetic pulse. In order to investigate the potential of this technology to classify cocoa percentages in chocolates, the terahertz spectra (0.5–10 THz) of five chocolate samples (50%, 60%, 70%, 80% and 90% of cocoa) were examined. The acquired data matrices were analyzed with the MATLAB 2019b application, from which the dielectric function was obtained along with the absorbance curves, and were classified by using 24 mathematical classification models, achieving differentiations of around 93% obtained by the Gaussian SVM algorithm model with a kernel scale of 0.35 and a one-against-one multiclass method. It was concluded that the combined processing and classification of images obtained from the terahertz time-domain spectroscopy and the use of machine learning algorithms can be used to successfully classify chocolates with different percentages of cocoa.

scholarly journals Effect of Children’s Weight Status on Physical Activity and Sedentary Behavior during Physical Education, Recess, and After School

2020 ◽  
Vol 9 (8) ◽  
pp. 2651
Author(s):  
Zachary C. Pope ◽  
Charles Huang ◽  
David Stodden ◽  
Daniel J. McDonough ◽  
Zan Gao
Keyword(s):  
Physical Activity ◽  
Body Mass Index ◽  
Elementary School ◽  
Physical Education ◽  
Sedentary Behavior ◽  
Body Mass ◽  
Weight Status ◽  
After School ◽  
Healthy Weight

Children’s body mass index may affect physical activity (PA) participation. Therefore, this study examined the effect of children’s weight status on underserved elementary school children’s PA and sedentary behavior (SB) throughout the segmented day. Participants were 138 children (X¯age = 8.14 years). Children’s height and weight were measured with subsequent classification of children as healthy weight or overweight/obese. Durations of moderate-to-vigorous PA (MVPA), light PA (LPA), and SB during physical education (PE), morning recess, lunch recess, after school, and overall were assessed via accelerometry over three days. Independent t-tests evaluated differences in children’s MVPA, LPA, and SB during each daily segment by weight status. Significantly higher MVPA was observed for children of healthy weight status versus children with overweight/obesity during morning recess, t(136) = 2.15, p = 0.03, after school, t(136) = 2.68, p < 0.01, and overall, t(136) = 2.65, p < 0.01. Interestingly, comparisons of children of healthy weight status and children with overweight/obesity’s LPA and SB during the after-school segment revealed a trend wherein children with overweight/obesity participated in slightly greater LPA/less SB than children of healthy weight status. Higher MVPA was observed among children of healthy weight versus children with overweight/obesity during most daily segments. Concerted efforts should focus on increasing MVPA among children with overweight/obesity.

scholarly journals Studying of Cultural Properties of Pyricularia oryzae Cav. Strains in the South of Russia

2021 ◽  
Vol 12 (1) ◽  
pp. 21-28
Author(s):  
Dmitry Nartymov ◽  
Evgeny Kharitonov ◽  
Elena Dubina ◽  
Sergey Garkusha ◽  
Margarita Ruban ◽  
...  
Keyword(s):  
Pyricularia Oryzae ◽  
Blast Disease ◽  
The South ◽  
Directional Growth ◽  
Rice Plants ◽  
Affected Plant ◽  
Cultural Traits ◽  
Plant Form ◽  
Using Data

This article presents the results of the development of a methodology for describing the main morphological and cultural traits of the Pyricularia oryzae Cav. strains widespread in the south of Russia. At the same time, the types of traits are identified and listed, which make it possible to unambiguously determine the uniqueness and variety of the pathogen. The relationships and patterns established using cluster and statistical analysis make it possible to identify the conditions for the development of a pathogen that determine its predominant forms. Thus, research shows that leaf forms of P. oryzae strains isolated from rice plants with leaf form of blast disease have an equally directional growth pattern of a colony with a felt structure, and strains isolated from neck-affected plant form often produce a zone of a colony with a clumpy structure. The classification of cultural traits will make it possible to obtain scientifically grounded and comparable data that can be used in the analysis of the interaction of P. oryzae strains with rice plants on various varieties and in various agro-technological conditions in order to improve and rationalize agricultural activities. The study opens up the possibility of using data in breeding, making it possible to identify forms of a pathogen that infect certain varieties.

Sign in / Sign up

Export Citation Format

Share Document