Diseño de un algoritmo para la avaluación de la actividad física i la intensidad del ejercicio basado en el smartphone
Year: 2014
Type: dissertation
Abstract: [ANGLÈS] The designed algorithm allows non-intrusive physical activity recognition using accelerometer data from the smartphone. Fourteen subjects of the same age (mean = 22 years) have used the Android application RSAF for a week in free conditions. The ReliefF and Sequential Forward Selection methods have been used for feature selection. Different Machine Learning models have been assessed for five users: Decision Tree, Naive Bayes and Support Vector Machine. In this work is also assessed the caloric expenditure of the physical activity. Accuracies up to 88.75% in predicted activities and up to 88.59% in caloric expenditure have been obtained. Finally, the results of the acquired information are presented in order to give a feedback of the user’s lifestyle.
Author: Carbonés Fargas, Bernat
Institution: Universitat Politècnica de Catalunya
Language: Catalan
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7113169865
