Optimizing Hybrid Feature Selection for Physical Activity Recognition Using Deep Learning on Wearable Sensor Data

Sharif, Rehan (2024) Optimizing Hybrid Feature Selection for Physical Activity Recognition Using Deep Learning on Wearable Sensor Data. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

This recognition of physical activities has recently received much attention because of its applications in health monitoring, sports analytics, and activity tracking. This research proposes a deep learning-based approach to optimize hybrid feature selection for physical activity recognition using the PAMAP2 Physical Activity Monitoring dataset. The main purpose of this research is to increase the precision and efficiency of activity recognition systems by leveraging spatial and temporal patterns within sensor data. A detailed pipeline is designed, starting with data preprocessing, then followed by hybrid feature selection by means of ElasticNet, Random Forest, and Mutual Information, and finally model development with CNN, LSTM, and a hybrid CNN-LSTM architecture. Each model is tested on the key metrics, and the LSTM model performed the best with an accuracy of 97.60%, precision of 97.59%, recall of 97.60%, and F1 score of 97.59%. The CNN model was the second best with an accuracy of 95.48%, precision of 95.52%, recall of 95.48%, and F1 score of 95.49%. The CNN-LSTM hybrid model achieved an accuracy of 95.62%, precision of 95.81%, recall of 95.62%, and F1 score of 95.55%. The selected features along with the best LSTM model are deployed through a Flask API, which enables real-time activity recognition from raw sensor data. It offers an innovative end-to-end activity recognition framework that incorporates hybrid feature selection techniques along with deep learning to build high levels of robustness and reliability in this end.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Singh, Jaswinder UNSPECIFIED
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning R Medicine > RA Public aspects of medicine > RA0421 Public health. Hygiene. Preventive Medicine > Personal Health and Hygiene
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	04 Sep 2025 14:52
Last Modified:	04 Sep 2025 14:52
URI:	https://norma.ncirl.ie/id/eprint/8803

Actions (login required)

View Item