Publication: Anomaly detection in human activity based on sensor data using machine learning
| datacite.subject.fos | oecd::Engineering and technology::Electrical engineering, Electronic engineering, Information engineering | |
| dc.contributor.author | Abd Rahman, Nur Masturina | |
| dc.date.accessioned | 2024-02-26T09:02:22Z | |
| dc.date.available | 2024-02-26T09:02:22Z | |
| dc.date.issued | 2022-07-01 | |
| dc.description.abstract | Machine Learning is the study of computer algorithms that can be improved automatically through experience and data analysis. Machine Learning is a branch of Artificial Intelligence where the systems can learn from data, identify patterns, and make decisions with minimal human intervention. Anomaly detection is a technique for identifying rare events in Machine Learning. There are a variety of domains applicable for anomaly detection such as system health detection, fault detection, event detection, detection, image, etc. The focus of these studies is to design and simulate anomaly detection of human activity by exploring the best techniques or algorithms for anomaly detection in Machine Learning. Three techniques have been chosen in this project that is DBSCAN, K-Means clustering, and Isolation Forest. This project is a software-based project which covers the development of the coding which contains anomaly detection with Machine Learning. The simulation is performed from two datasets that is human activity based on sensor data which are heart rate, body temperature, BMI, and accelerometer. The three techniques are applied to a dataset and the result is compared. From the result, the accuracy of every method for each dataset has been calculated. The average accuracy for method DBSCAN provided around 70% of accuracy, K-Means provided around 82% accuracy and Isolation Forest provided an accuracy of around 90%. It shows that the Isolation Forest performs quite well in detecting anomalies in this project. The comparison shows that the best techniques are Isolation Forest. | |
| dc.identifier.uri | https://erepo.usm.my/handle/123456789/18464 | |
| dc.language.iso | en | |
| dc.title | Anomaly detection in human activity based on sensor data using machine learning | |
| dc.type | Resource Types::text::report | |
| dspace.entity.type | Publication | |
| oairecerif.author.affiliation | Universiti Sains Malaysia |