TY - GEN
T1 - Enhancing Classification of Elderly Fall Detection System using Tuned RBF-SVM
AU - Miawarni, Herti
AU - Sardjono, Tri Arief
AU - Setijadi, Eko
AU - Gumelar, Agustinus Bimo
AU - Wijayanti,
AU - Purnomo, Mauridhi Hery
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - An elderly FDS or Fall-related incident Detection System can reduce the severity of late treatment after a falling incident by shortening the period between the inevitable fall and the proper medical care. A failure to comply the need of urgent medical attention to the falling victims can lead to painful death. But first, a machine must be trained to automatically and accurately detect a falling incident. The Support Vector Machine (SVM) is generally acknowledged as one of the best tree-based machine learning algorithms for classification and regression data analysis. This tree-based Deep Learning model has been used in fields, together with bioinformatics, face recognition, and image recognition, and urgent cases such as FDS. This paper defines a work of FDS using SVM. We used 16,261 instances and 33 attributes of fall simulation from eHomeSeniors dataset, in which the dataset employs the infrared thermal sensor. This dataset stated more than 15 classes of fall, to achieve best result in capturing falling incidents. We also set the gamma value from the default 0.01, to 0.9, with no normalization or standardization needed in the process. The training data and testing data are also split into the scale of 50:50, 60:40, 70:30, 80:20, and 90:10. Consequently, we were able to reached 84.62% accuracy and 50.32 seconds learning runtime at data split of 90:10.
AB - An elderly FDS or Fall-related incident Detection System can reduce the severity of late treatment after a falling incident by shortening the period between the inevitable fall and the proper medical care. A failure to comply the need of urgent medical attention to the falling victims can lead to painful death. But first, a machine must be trained to automatically and accurately detect a falling incident. The Support Vector Machine (SVM) is generally acknowledged as one of the best tree-based machine learning algorithms for classification and regression data analysis. This tree-based Deep Learning model has been used in fields, together with bioinformatics, face recognition, and image recognition, and urgent cases such as FDS. This paper defines a work of FDS using SVM. We used 16,261 instances and 33 attributes of fall simulation from eHomeSeniors dataset, in which the dataset employs the infrared thermal sensor. This dataset stated more than 15 classes of fall, to achieve best result in capturing falling incidents. We also set the gamma value from the default 0.01, to 0.9, with no normalization or standardization needed in the process. The training data and testing data are also split into the scale of 50:50, 60:40, 70:30, 80:20, and 90:10. Consequently, we were able to reached 84.62% accuracy and 50.32 seconds learning runtime at data split of 90:10.
KW - Elderly
KW - Fall Detection System
KW - Infrared Thermal Sensor
KW - Support Vector Machine
KW - eHomeSeniors
UR - http://www.scopus.com/inward/record.url?scp=85135884403&partnerID=8YFLogxK
U2 - 10.1109/IST55454.2022.9827716
DO - 10.1109/IST55454.2022.9827716
M3 - Conference contribution
AN - SCOPUS:85135884403
T3 - IST 2022 - IEEE International Conference on Imaging Systems and Techniques, Proceedings
BT - IST 2022 - IEEE International Conference on Imaging Systems and Techniques, Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2022 IEEE International Conference on Imaging Systems and Techniques, IST 2022
Y2 - 21 June 2022 through 23 June 2022
ER -