The Impact of Data Reduction on Wearable-Based Human Activity Recognition

Nowadays, human activity recognition (HAR) plays an important role in wellness-care and context-aware systems. Human activities can be recognized in real-time by using sensory data collected from various sensors built in smart mobile devices. Recent studies have focused on HAR that is solely based on triaxial accelerometers, which is the most energy-efficient approach. However, such HAR approaches are still energy-inefficient because the accelerometer is required to run without stopping so that the physical activity of a user can be recognized in real-time. In this paper, we propose a novel approach for HAR process that controls the activity recognition duration for energy-efficient HAR. We investigated the impact of varying the acceleration-sampling frequency and window size for HAR by using the variable activity recognition duration (VARD) strategy. We implemented our approach by using an Android platform and evaluated its performance in terms of energy efficiency and accuracy. The experimental results showed that our approach reduced energy consumption by a minimum of about 44.23% and maximum of about 78.85% compared to conventional HAR without sacrificing accuracy.

Download Full-text

A Comparative Analysis of the Impact of Features on Human Activity Recognition with Smartphone Sensors

Proceedings of the 23rd Brazillian Symposium on Multimedia and the Web - WebMedia '17 ◽

10.1145/3126858.3126859 ◽

2017 ◽

Cited By ~ 9

Author(s):

Wesllen Sousa ◽

Eduardo Souto ◽

Jonatas Rodrigres ◽

Pedro Sadarc ◽

Roozbeh Jalali ◽

...

Keyword(s):

Comparative Analysis ◽

Activity Recognition ◽

Human Activity ◽

Human Activity Recognition ◽

Smartphone Sensors ◽

The Impact

Download Full-text

Evaluating the Impact of a Two-Stage Multivariate Data Cleansing Approach to Improve to the Performance of Machine Learning Classifiers: A Case Study in Human Activity Recognition

Sensors ◽

10.3390/s20071858 ◽

2020 ◽

Vol 20 (7) ◽

pp. 1858 ◽

Cited By ~ 1

Author(s):

Dionicio Neira-Rodado ◽

Chris Nugent ◽

Ian Cleland ◽

Javier Velasquez ◽

Amelec Viloria

Keyword(s):

Machine Learning ◽

Activity Recognition ◽

Human Activity ◽

Human Activity Recognition ◽

Training Process ◽

Quality Of Data ◽

Machine Learning Classifiers ◽

Learning Classifiers ◽

The Impact

Human activity recognition (HAR) is a popular field of study. The outcomes of the projects in this area have the potential to impact on the quality of life of people with conditions such as dementia. HAR is focused primarily on applying machine learning classifiers on data from low level sensors such as accelerometers. The performance of these classifiers can be improved through an adequate training process. In order to improve the training process, multivariate outlier detection was used in order to improve the quality of data in the training set and, subsequently, performance of the classifier. The impact of the technique was evaluated with KNN and random forest (RF) classifiers. In the case of KNN, the performance of the classifier was improved from 55.9% to 63.59%.

Download Full-text

AI Assisted Human Activity Recognition (HAR)

International Journal of Engineering and Advanced Technology - Regular Issue ◽

10.35940/ijeat.f8575.088619 ◽

2019 ◽

Vol 8 (6) ◽

pp. 2143-2148

Keyword(s):

Machine Learning ◽

Activity Recognition ◽

Human Activity ◽

Human Activity Recognition ◽

Supervised Machine Learning ◽

Learning Models ◽

Evaluation Metric ◽

Gyroscope Sensor ◽

The Impact ◽

Machine Learning Models

Human Activity Recognition and assisting user on the basis of his context is attracting researchers since decade Researchers are working in the area to increase the accuracy of detection by various means. The challenging issue is to determine the correct supervised classifier for the detection purpose. This paper intent to examine the methodology used to recognize HAR and the impact of classifiers practiced in training and Testing. We have also tried to identify the suitable supervised machine learning model for HAR. Data of 30 Users with 561 features belonging to accelerometer and gyroscope sensor of smartphone from UCI repository is used for evaluation purpose. Nine different supervised machine learning Models are trained and tested on the dataset. The result concludes that HAR is a process which depends upon the classifiers used. It also conclude that out of 9 different Machine learning models ANN performs well and after that SVM, kNN, Random Forest and Extra Tree are equally good models for the purpose of HAR with Accuracy and execution time as the performance evaluation metric.

Download Full-text

Analyzing the impact of different action primitives in designing high-level human activity recognition systems

Journal of Ambient Intelligence and Smart Environments ◽

10.3233/ais-130223 ◽

2013 ◽

Vol 5 (5) ◽

pp. 443-461 ◽

Cited By ~ 1

Author(s):

Atif Manzoor ◽

Hong-Linh Truong ◽

Alberto Calatroni ◽

Daniel Roggen ◽

Mélanie Bouroche ◽

...

Keyword(s):

Activity Recognition ◽

Human Activity ◽

Human Activity Recognition ◽

Recognition Systems ◽

High Level ◽

The Impact

Download Full-text

Human Activity Recognition: A Comparative Study to Assess the Contribution Level of Accelerometer, ECG, and PPG Signals

Sensors ◽

10.3390/s21216997 ◽

2021 ◽

Vol 21 (21) ◽

pp. 6997

Author(s):

Mahsa Sadat Afzali Afzali Arani ◽

Diego Elias Costa ◽

Emad Shihab

Keyword(s):

Activity Recognition ◽

Human Activity ◽

Inertial Sensors ◽

Human Activity Recognition ◽

Daily Activities ◽

Random Forest Models ◽

Correlation Based Feature Selection ◽

Feature Selection Approach ◽

Electrocardiogram Ecg ◽

The Impact

Inertial sensors are widely used in the field of human activity recognition (HAR), since this source of information is the most informative time series among non-visual datasets. HAR researchers are actively exploring other approaches and different sources of signals to improve the performance of HAR systems. In this study, we investigate the impact of combining bio-signals with a dataset acquired from inertial sensors on recognizing human daily activities. To achieve this aim, we used the PPG-DaLiA dataset consisting of 3D-accelerometer (3D-ACC), electrocardiogram (ECG), photoplethysmogram (PPG) signals acquired from 15 individuals while performing daily activities. We extracted hand-crafted time and frequency domain features, then, we applied a correlation-based feature selection approach to reduce the feature-set dimensionality. After introducing early fusion scenarios, we trained and tested random forest models with subject-dependent and subject-independent setups. Our results indicate that combining features extracted from the 3D-ACC signal with the ECG signal improves the classifier’s performance F1-scores by 2.72% and 3.00% (from 94.07% to 96.80%, and 83.16% to 86.17%) for subject-dependent and subject-independent approaches, respectively.

Download Full-text