Data Analysis and Classification of mHealth Shimmer2 Sensor Data sets for Human Physical Activities Recognition
Maria Navin J R1, Sridevi N2, Pankaja R3
1Maria Navin J R, Department of ISE, Sri Venkateshwara College of Engineering, Bangalore (Karnataka), India.
2Sridevi N, Department of CSE, Sri Venkateshwara College of Engineering, Bangalore (Karnataka), India.
3Pankaja R, Department of ISE, Sri Venkateshwara College of Engineering, Bangalore (Karnataka), India.
Manuscript received on 22 August 2019 | Revised Manuscript received on 03 September 2019 | Manuscript Published on 16 September 2019 | PP: 844-848 | Volume-8 Issue-2S6 July 2019 | Retrieval Number: B11560782S619/2019©BEIESP | DOI: 10.35940/ijrte.B1156.0782S619
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Human physical activity recognition is one of the important attributes for well being of human health and activity recogni-tion field. A physical signal with respect to time based directed toward recognizes human activity events is projected, in the view of this work usage wearable accelerometer sens-ing devices were implanted taking place the human body lo-cation on the upper body Chest Sensor (CS), Left Ankle Sen-sor (LAS) and Right Lower Arms Sensor (RLAS). Accelerome-ter feature extracted based acceleration signals with respect to time, physical appearance of the accelerometer x, y, and z dimension values reported/recorded using shimmer2 weara-ble sensor device is recommended at the categorization of the 10 different users was performed 12 different types human activities, including vigorous and moderate activities. User ages between 24 to 29 years and human body weight (HBW) are 53 to 83 Kg=m2. Results were on view a large validity per-formance precision and recall were getting 95 for each human activities. The whole classifiers accuracy results for all combi-nation of the feature set of all sensors is 98%. The considered work could be used to observe the human body motion of different body location of users and to perform data analysis and classification.
Keywords: Data Analysis, Activity Recognition, Wearable Sensor, Physical Activity, Classification.
Scope of the Article: Data Analysis