The study for score classification in berg balance scale using Inertial sensors

초록

In this paper, a the Berg Balance Scale(BBS) scoring algorithm using inertial sensors was studied. Forty-eight patients with brain disease between the ages of 50 and 80 participated in the experiment. Inertial snesors were attached to the patient's body and BBS operating data were obtained. Multi-layer perceptron (MLP) machine learning algorithms were used for scoring. In feature extraction process, sensor data with a high contribution to scoring were selected and the amplitudes of frequency components from 0.1 to 15 Hz were used. The classification accuracy was 88% for the MLPmodel.