Using Time Series to Determine Model Rank with a Practical Application

Abstract

A time series is a set of sequential data over a specific period of time and is used to analyze, understand, and describe variables over time. It is used to predict future phenomena and make decisions. Time series analysis relies on advanced mathematical statistics such as prediction and estimation models. This helps you understand the phenomenon over a specific period of time and improve planning and effective decision-making in different contexts. In this research, the ARMAX model was used, which represents an autoregressive moving average model with external inputs. It is a type of equation error model belonging to the black box models. This was done by studying crude oil prices and the price of the dollar to choose the best rank for the ARMAX model using the statistical criteria (AIC, AICC, BIC) after testing the stationarity of the time series.