Fall detection using machine learning algorithms

In this paper, the recognition of and the differentiation between fall activities and activities of daily living (ADL) was performed using the MobiFall dataset. A large database was constructed to train and validate the model. Feature selection methods were implemented to reduce dimensionality. Five different classification algorithms were implemented and evaluated based on their accuracy' sensitivity, and specificity achieved. The k-Nearest Neighbors' algorithm obtained an overall accuracy of 87.5% with a sensitivity of 90.70%, and a specificity of 83.78%.