Recognition physical activities with optimal number of wearable sensors using data mining algorithms and deep belief network

Publication Type:
Conference Proceeding
Citation:
Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS, 2017, pp. 2871 - 2874
Issue Date:
2017-09-13
Metrics:
Full metadata record
Files in This Item:
Filename Description Size
EMBC2017-2_final.pdfAccepted Manuscript version214.19 kB
Adobe PDF
© 2017 IEEE. Daily physical activities monitoring is benefiting the health care field in several ways, in particular with the development of the wearable sensors. This paper adopts effective ways to calculate the optimal number of the necessary sensors and to build a reliable and a high accuracy monitoring system. Three data mining algorithms, namely Decision Tree, Random Forest and PART Algorithm, have been applied for the sensors selection process. Furthermore, the deep belief network (DBN) has been investigated to recognise 33 physical activities effectively. The results indicated that the proposed method is reliable with an overall accuracy of 96.52% and the number of sensors is minimised from nine to six sensors.
Please use this identifier to cite or link to this item: