Question

Why deseasonalize and seasonalize for forecasting are used?

Answer 1

Why deseasonalize and seasonalize for forecasting are used?

1. Forecasting usually involves use of past data to predict the future trends or data.
   • For example, a autombile manufacturer may need to predict sales for next year to decide the production requirements and to balance capacity and demand.
2. Past data used for predicting may have some existing inherent patterns within the data.
3. These patterns may be, gennerally, in the form of trends or seasonality.
   • For example, there may be an continuously increase in sales of automobiles from year to year. This is the overall trend component.
   • At the same time, the automobile sales within a year will have variations from month to month or quarter to quarter. This is an example of the seasonal component of data.
4. Deseasonalising of forecasts is done to separate and identify the trend component of the past data from the seasonal variations.
5. Once, the trend component is separated and a forecast based on this trend is generated, we want to know how the data will vary from each season to season within the forecast period(could be month to month, quarter to quarter or eaven within days of the week like weekdays and weekends).
6. Seasonalising the forecast is thus done on the deseasonalised forecast to identify the variations from one season to another season within the forecasted period.

So, For example,

In the automobile manufacturer case, we will first deseasonalise data to identify whether there is any increasing or decreasing trend in sales and use it to predict the demand for next year. Then the predicted demand will be seasonalised to identify how the sales will vary from one month to other month within that year.