Title: Classification of human activity recognition using smartphones

Abstract: Smartphones have been the most popular and widely used devices among means of communication. Nowadays, human activity recognition is possible on mobile devices by embedded sensors, which can be exploited to manage user behavior on mobile devices by predicting user activity. To reach this aim, storing activity characteristics, Classification, and mapping them to a learning algorithm was studied in this research. In this study, we applied categorization through deep belief network to test and training data, which resulted in 98.25% correct diagnosis in training data and 93.01% in test data. Therefore, in this study, we prove that the deep belief network is a suitable method for this particular purpose.