Elderly people are those who are 60 years of age or older. Elderly people are more likely to fall due to age-related reductions in physiological processes, particularly bone and muscular functions. Falling is one of the symptoms that can be lethal. These effects may cause mortality if prompt medical assistance is not received due to the deterioration in numerous organ functions required to maintain body homeostasis. Previous studies have tested the Random Forest model from acceleration and gyroscope measurements to identify human activity. In this research, feature extraction was carried out utilizing variables such as maximum, minimum, mean, median, kurtosis, skewness, and variance that were obtained from the Acceleration data from accelerometer sensor in type IMU LSM9DS1. The Acceleration data include Acceleration X, Acceleration Y, Acceleration Z and Acceleration Magnitude. To evaluate the effectiveness of the Decision Tree model, cross-validation will be employed. The best feature extraction values were Magnitude Acceleration Variance, X Acceleration Maximum, Magnitude Acceleration Maximum, Z Acceleration Median, and Z Acceleration Variance, with a Decision Tree model accuracy rate of 99.8%. Key Words: Elderly, Fall, Tendency to Fall, Decision Tree, Statistical Feature
Copyrights © 2023