Online DCA for Times Series Forecasting Using Artificial Neural Network
Abstract
In this work, we study the online time series forecasting problem using artificial neural network. To solve this problem, different online DCAs (Difference of Convex functions Algorithms) are investigated. We also give comparison with online gradient descent—the online version of one of the most popular optimization algorithm in the collection of neural network problems. Numerical experiments on some benchmark time series datasets validate the efficiency of the proposed methods.