“…In this case, there is no need to consider the causes or relationships underlying changes in the data. Common univariate time series forecasting methods include the moving average method (Armstrong, 1985), the exponential smoothing method (Fildes & Lusk, 1984), the Box-Jenkins method (Hill & Fildes, 1984), the ARARMA model (Meade & Smith, 1985), the Pandit-Wu method (Pandit & Wu, 1983), the intervention analysis model (Thury & Anderson, 1980), the state space model and the Bayesian forecasting method (Abraham & Ledolter, 1983). Instead, multivariate forecasting focuses on analyzing causal or correlation relationships between two or more variables.…”