ABSTRACT
Many products have seasonal or cyclic demand patterns. These can be caused by natural seasonal factors – for example, temperature, rainfall or the probability of thunder storms – or by other seasonal factors – such as holiday patterns, Christmas or other religious festivals, or financial year-ends. These factors can be the overriding cause of demands, or small contributions for fine-tuning the forecast. Forecasting seasonal demand requires at least a year's worth of history or an estimate. Before the forecasting can start, seasonal demand factors have to be established and an initial forecast made for the average demand. The classic technique for calculating seasonal demand is base series forecasting. Base series produces seasonal demand forecasts as an extension to exponential smoothing. Exponential forecasting techniques have proved to be successful in ensuring stock predictions are good in practical situations which generally suffer from poor data integrity, feedback from the market and changing demand patterns.
