New
Product Forecasting On-line:
Feature 2: Forecasting with
Some History
When some product history is available for the new product, then a combination of
statistical techniques and assumptions about the product's long term saturation
level can be used to produce forecasts. In order to use this module, press the browse
button and read in your data in a one column text file as shown in the example file
called SDATA1.TXT. This file can be downloaded to your hard drive and read back
into Feature 2's forecasting module. The one requirement is that the last data point
must be greater than any previous datapoint in your product's history. As a default,
the long run saturation level is assumed to be equal to five times the value of
the last datapoint. The user then is encouraged to put in their own assumptions
so that regression estimates can be developed representing
a logistic diffusion process. The procedure also performs an automatic grid search
for the optimal saturation level with a default range 25% above and below the user's
suggested saturation level.
However, results from this feature should be used with caution.