Our partner Business Forecast Systems, Inc.(BFS, Inc) has announced that Forecast Pro TRAC Version 7 will be released shortly ! It introduces 3 important new functionalities:
- Machine Learning: integrates new cutting-edge methodologies to generate more accurate and reliable forecasts ;
- Product Mapping: allows to easily handle product and warehouse introductions and transitions ;
- Forecast Value Add: allows to better track accuracy across each step of the forecasting process.
According to our partner, additional software improvements include the ability to export external rows to Excel collaboration sheets, add item-specific calculation rows to the override grid, seamlessly open and convert projects from different locale, recover unsaved projects, and much more.
Machine Learning
- With Machine Learning (ML) powered forecasting, producing reliable forecasts and improving accuracy is easier than ever. Either automatically the user can apply ML using the Expert Selection algorithm, or choose Machine Learning as a new method for creating customized models.
- Forecast Pro TRAC V7 integrates Extreme Gradient Boosting (XG Boost) Decision Trees, a Machine Learning technique proven to be best-in-class in forecasting competitions. XG Boost is readily accessible to all Forecast Pro users by leveraging the same AI-driven algorithm that powers Expert Selection to select and train the XG Boost models.
- Automatically applying Machine Learning using only historic data is simple with the Expert Selection algorithm. Creating customized models with user-specified features, such as event schedules or explanatory variables, is fast and easy when the user chooses Machine Learning as forecasting method.
Product Mapping
- To generate a good forecast, the user needs historical data. However, he/she may have limited history in case of launching a new product or introducing an existing product in a new location.
- Product Mapping (aka Item Supersession) allows to create a “forecast history” by mapping history from other similar items into the item that the user wants to forecast. This forecast history will then be used to generate a forecast.
- The user can define Product Mappings in Forecast Pro via a user-friendly interface or through file(s).
Forecast Value Add Reporting
- Forecast Value Add (FVA) Reporting helps refining the forecasting process by identifying which steps or participants in the forecasting process are adding value and which are making your forecasts less accurate.
- Global FVA reports helps quickly identifying which items and which process steps are adding or losing the most value. Item-specific FVA reports provide an in-depth look at specific end items or groups.
- FVA measures the accuracy of each override versus the statistical forecast or any other prior override