Predicting fall in elderly people using machine learning

Fall is a serious health problem, it may threaten the life of many people in general and the life of the elderly in particular. That is why we tried very hard to develop a system to notify the mechanisms of their fall from the lane by monitoring their movement by means of wearable sensors in certain places on their bodies. This paper presents five supervised machine learning algorithms (SVM, Neural Network, Decision Tree, Random Forest, and Naïve Bayes) to predict fifteen falls in the elderly. We compared the five models in terms of performance measures (accuracy, precision and recall), and the Random Forest model achieved the best result with an accuracy of 95.91%. In the future, we plan to improve the results by pre-processing the data for better features and then classifying and predicting them using advanced algorithms in artificial intelligence techniques to build a system that will be able to predict with high speed and accuracy.