Using Recent Observed Data to Improve Forecasts

7 Using Recent Observed Data to Improve Forecasts 

How can the approaches described earlier be improved if there is access to more current data regarding season demands One approach is to use Bayesian updating ([35]) to improve demand forecasts. To understand this approach, consider the classification of products and their associated actual demand distribution as in the example in the previous section. Before discussing details of the approach, consider how collected data can be used to project season demand for a product. 

As an example, suppose that a product had been classified as a runner, but the early demand projection suggested a demand level of 25 units. Then the Bayesian updated weights for that product are obtained as follows  

Using the data provided earlier, the calculations would be / Runner 0.0251 X 0.65 

---