Comparison of Algorithm Selection to Analyze Elderly Activity Recognition Based on Sensor Data Using R Program

This paper presents a comparison of algorithms used to classify human activity of elderly by sensor data from UCI Machine Learning Repository. We compare three popular algorithms used to classify activities as Artificial Neural Networks, Support Vector Machine and C4.5 Decision Tree to find the most efficient algorithm for classifying human activities. This research used data set “Activity recognition with healthy older people using a batteryless wearable sensor Data Set” of 6000 records. The result is that Artificial Neural Networks has the highest classification accuracy of 96.5%, followed by the Support Vector Machine 96.41% and the C4.5 Decision Tree 95.75%. These experimental data can be applied to the system of detecting or monitoring the activities of the elderly.