The Use of Artificial Neural Network in Time Series Forecasting

• Activity Number: 41 - Statistical Analysis of Networks
• Type: Contributed
• Date/Time: Sunday, July 30, 2017 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Learning and Data Science
• Author(s): Taysseer Sharaf*
• Companies: University of Michigan- Dearborn
• Keywords: Time series forecasting ; Neural Networks ; Statistical learning ; Bayesian learning

Abstract

In the current study we explore the use of artificial neural network in time series forecasting. We discuss the use of recursive neural network in time series modeling, and how to choose the best neural network structure that best fits the data. In addition, we represents the use of Bayesian learning to train the neural network and how it affects the prediction accuracy. We also validate and compare neural network model with the regular seasonal ARIMA model.