Some values in your historical or related data may skew your forecasts. Examples for such values include outliers and zeros that represent missing data and aren't true zeros.

The exclude values functionality can replace undesired values with values that fit better with your time series data. The new values are generated with the use of an automated logic and are based on where the values are in the time series.

To use the exclude values functionality, follow the instructions detailed in Step 1: Create Forecaster modules. Then, return to this page and follow the instructions below.

Note: The exclude values functionality doesn't impact your original data, but impacts how Forecaster interprets your data.


To use the exclude values functionality:

  1. Open the historical or related data module that contains the values you want to replace.
  2. Add a new numeric line item to the relevant module, and give it the same name as the original line item. Then, add the string ___exclude_value  to the end of the name of this new line item.
  3. Make sure that this line item is included in the saved view, from the data collection step.

To name the new line item, follow this format:

Original line item name + three underscores + exclude + one underscore + value


Example:
If you have a line item called Units sold, the new line item would be Units sold___exclude_value:

  1. Identify the forecast items with the values you want to exclude, and insert 1 to the new exclude value line item. All other values remain as 0. 

The table below shows a grid of historical data with excluded values. The excluded values are in bold italics.

Historical data




Units soldUnits sold___exclude_value

Item 100

Item 200

Item 300

Item 410

Item 500

Item 600
Week 1 FY20Item 701

Item 800

Item 9841

Item 1000

Item 1100

Item 1200

Item 1300

Item 1400

Item 1500

Item 1610

Item 1710

Item 1800