We describe an artificial neural network (ANN) for analyzing damped oscillations in a simple pendulum system by using a machine learning (ML) algorithm. We have first shown how to construct a simple ANN consisting of three layers-input, hidden and output, with each layer being composed of neurons representing a relevant feature of the oscillating pendulum. The train and test datasets for the ANN have been taken from the experimental data collected by using the methodology of a previously communicated work. A ML optimization algorithm called stochastic gradient descent has been employed in the neural network to predict the type of pendulum according to the values of the mass, size and damping coefficient of the pendulum.