The identification of past patterns and their extrapolition into the future is to treat them as comprising four elements’ seasonality, trend, cyclicality and randomness (error terms).

The first three elements, seasonality, trends and cyclicality can be extrapolated to prepare more accurate forecast, while the randomness cannot be predicted. However, randomness indicates the extent of variation between the actual and predicted results. Randomness is also called residuals. Randomness can help us determine the extent of uncertainty in our future predictions.