Activities of Daily Living Classification using Recurrent Neural Networks

In this paper we address the problem of classifying the daily life activities of a person out of sensor based monitored data. We propose the use of recurrent neural networks to track of successive sensor data inputs and Long Short-Term Memory cells to address the issues regarding the long-time dependencies in activities' monitored data. The recurrent neural network model was implemented using TensorFlow library. The results are promising showing a mean accuracy of 82.5 using basic cross validation respectively 87.16% using leave one subject out method. Our results are comparable with the ones reported in the state of the art being slightly better in case of the leave one person out validation approach.