Forecasting of Demand Using Exponential Smoothing

Data Forecasting and Exponential Smoothing

Data forecasting is predicting future values using historical values. There are many methods of forecasting, in this paper, we use three forecasting methods: exponential smoothing. The forecasting method is useful when the data shows the non-linear trend. If the data does show seasonality, moving averages might be used, but when the stock data has no linear trend, then exponential smoothing is used (Hyndman, et al., 2012) The data used is the time series demand data for Oil demand in Saudi Arabia which would in effect determine the amount of oil to be processed and exported in future. The data was selected to give a case study on how exponential smoothing can be applied. Other methods such as regression analysis can only be used in forecasting between the dependent and independent variables, and therefore with data without any relationship such as time-based series demand data presented, regression analysis cannot be applied. Moving averages also need more data such as sample population size which is not present in the data being used in this example.

The Additive Method in Forecasting

Another forecasting tool which is useful but could not be applied to the data is the additive method which assumes that the data that depend on the time series additively. One example of data which additive method could be used is the monthly data, the method is useful and could also apply to the data although exponential smoothing seemed more attractive in predicting the time series data.

Recommendations based on the Data

The recommendations I would give from the data is that within the next two or three weeks, the demand would remain at slightly above 255,000, with a slight increase. The company should, therefore, produce more goods and retain the current manufacturing capacity with a small increase in production (Taylor, 2013). Despite the fluctuation, the forecast of the demand remains reliable high above the current figures at 255,610 for the next three weeks.

References

Hyndman, R. J., Koehler, A. B., Snyder, R. D., & Grose, S. (2012). A state space framework for automatic forecasting using exponential smoothing methods. International Journal of forecasting, 18(3), 439-454.

Taylor, J. W. (2013). Short-term electricity demand forecasting using double seasonal exponential smoothing. Journal of the Operational Research Society, 54(8), 799-805.