More accurate AI Forecasting on bulk at detailed level

The AI Forecasting can work at a consolidated level however, its not accurate at the detailed level because it evenly spreads. When AI forecasting at the consolidated level, we would like it to have the option to perform the AI forecasting routine for each of the individual lines encompassed within that in turn. Current workaround is to select each individual child row one at a time and run the forecast routine for each one in turn which is too cumbersome when wanting to perform this routine on mass for lots of customers and products in one go.

E.g. if AI forecasting is applied to a consolidated level with 10 customers, and the forecast amount was £1000, it would apply £100 to each of the 10 customers even if some customers clearly purchase higher amounts than others. To get the AI forecasting to work in the correct proportion for each customer, you would need to do individual AI forecasting predictions for each of the 10 customers individually. I would like the option for the routine to perform the AI forecasting on bulk for each individual line rather than evenly spread the number.

Needed By

Yesterday (Let's go already!)

Post comment

Guest

Reply
| Mar 3, 2022

In addition, we're providing options for proportional or relative proportional data spreading to a forecast as of PAW 2.0.66.
Benefit – allow to choose data spreading for more accurate forecasting

Proportional data spreading distributes a specified value among cells proportional to existing cell values.
Relative proportional data spreading distributes values to the leaves (children) of a consolidation proportional to the leaves of a reference cell or cells. This is useful when you want to create a forecast based on a known previous time period.
Both of these spreading methods can be applied to a forecast.

0 reply Hide replies

Guest

Reply
| Sep 7, 2021

Thank you for the feedback.

1. A new option was added (in PA Workspace 66) to provide more spreading options for forecasts at the consolidated level. The forecast side-panel now includes options to spread the forecast value by a) the previous period (last month) or b) the equivalent historical period based on seasonality (same month last year).
2. I've associate this idea with a roadmap item for creating a forecast baseline or forecasting multiple items in-bulk. The feature will allow the planning analyst to define and execute and large number of algorithmic forecasts (thus seeding algorithmic values for items on multiple TM1 slices).
3. A small point for your infomation, today the user is able to select multiple items to forecast form a exploration view. I realize this is limiting for larger use-cases.
Thanks again.

0 reply Hide replies

By clicking the "Post Comment" or "Submit Idea" button, you are agreeing to the IBM Ideas Portal Terms of Use.
Do not place IBM confidential, company confidential, or personal information into any field.

Please enter your email address

RELATED IDEAS

More accurate AI Forecasting on bulk at detailed level