US 12,340,445 B1
Hierarchical modeling node for visual forecasting
Michele Angelo Trovero, Cary, NC (US); Mahesh Vijaykumar Joshi, Cary, NC (US); Steven Christopher Mills, Raleigh, NC (US); Phillip Mark Helmkamp, Apex, NC (US); Youngjin Park, Durham, NC (US); Iman Vasheghani Farahani, Charlotte, NC (US); Rajib Nath, Pune (IN); Kritika Misra, Maharashtra (IN); Vilochan Suresh Muley, Pune, IN (US); and Ran Bi, Apex, NC (US)
Assigned to SAS INSTITUTE INC., Cary, NC (US)
Filed by SAS Institute, Inc., Cary, NC (US)
Filed on Dec. 13, 2024, as Appl. No. 18/980,190.
Claims priority of provisional application 63/685,854, filed on Aug. 22, 2024.
Claims priority of application No. 202411063211 (IN), filed on Aug. 21, 2024.
Int. Cl. G06T 11/20 (2006.01); G06F 3/0482 (2013.01); G06F 8/34 (2018.01); G06F 8/38 (2018.01)
CPC G06T 11/206 (2013.01) [G06F 3/0482 (2013.01); G06F 8/34 (2013.01); G06F 8/38 (2013.01)] 30 Claims
OG exemplary drawing
 
1. A system comprising:
one or more processors; and
one or more memories including program code that is executable by the one or more processors to perform operations including:
generating a graphical user interface (GUI) for a piece of forecasting software, wherein the GUI includes a drag-and-drop canvas comprising a set of graphical nodes arranged to define an overall forecasting pipeline, each node in the set of graphical nodes corresponding to a respective operation in the overall forecasting pipeline;
based on a user input, attaching a hierarchical modeling node to the set of graphical nodes on the drag-and-drop canvas, wherein the hierarchical modeling node enables a user to define a time series hierarchy comprising a plurality of levels, and wherein the hierarchical modeling node enables one or more users to customize separate level pipelines for each level of the time series hierarchy independently of the other level pipelines for the other levels of the time series hierarchy, each of the level pipelines being a respective subpart of the overall forecasting pipeline;
executing the level pipelines for the plurality of levels of the time series hierarchy to generate a plurality of forecasts, each forecast of the plurality of forecasts corresponding to a respective level of the time series hierarchy;
executing a reconciliation process on the plurality of forecasts to generate a plurality of reconciled forecasts for the plurality of levels of the time series hierarchy, each reconciled forecast of the plurality of reconciled forecasts corresponding to a respective forecast of the plurality of forecasts; and
generating a visualization of a reconciled forecast of the plurality of reconciled forecasts.