Activity recognition is considered as a vital assignment in numerous applications, especially in medicinal services administrations. Among these applications incorporate restorative symptomatic, checking of clients’ every day schedule and discovery of strange cases. This paper presents a methodology for the action acknowledgment utilizing an accelerometer sensor installed in a cell phone. This methodology utilizes an openly accessible accelerometer dataset as the crude info flag. The highlights of the flag are chosen in view of the time and recurrence area. At that point, Principal Component Analysis (PCA) is used to diminish the dimensionality of the highlights and concentrate the most huge ones that can characterize human exercises. A correlation process is performed between the first crude information also, PCA-based highlights and moreover, time and recurrence space highlights are likewise analyzed utilizing a few machine learning classifiers.