Dimensionality Reduction Applied to Time Response of Linear Systems Using Autoencoders

An Autoencoder is a multilayer neural network that is used as a powerful tool to perform dimensionality reduction. Due to its structure, is possible to find low-dimensional representations of high-dimensional data in its most hidden layer. In this paper, a deep autoencoder is used to perform a compact representation of the time response of linear systems. The parameters of deep autoencoder are trained using gradient descent and backpropagation. The proposed method is validated with five different first-order systems and six of second order. Results show that is possible to use a deep autoencoder to capture the dynamical behavior of a dynamical system in its latent layer.