US 12,260,351 B2
Method and system for predicting the probability of regulatory compliance approval
Jonathan Levine, Rochester, NY (US); Ray Uri Merriam, Rochester, NY (US); Stephen Kyle Korndoerfer, West Henrietta, NY (US); Larry Glass, Pittsford, NY (US); Michael Wiseman, Webster, NY (US); Vijayakrishna Nama, Webster, NY (US); and Joseph Martin St. Germaine, Fairport, NY (US)
Assigned to Xerox Corporation, Norwalk, CT (US)
Filed by Xerox Corporation, Norwalk, CT (US)
Filed on Nov. 2, 2018, as Appl. No. 16/179,051.
Prior Publication US 2020/0143277 A1, May 7, 2020
Int. Cl. G06N 7/01 (2023.01); G06Q 10/08 (2024.01)
CPC G06N 7/01 (2023.01) [G06Q 10/08 (2013.01)] 14 Claims
OG exemplary drawing
 
1. A method for automatically predicting a probability of regulatory compliance approval based on data contained in a data structure, comprising:
configuring said data structure comprising first data collated and collected from a plurality of regulators and a plurality of value chain participants, said data structure provided by an RCS (Regulatory Compliance System) that collates the first data and restricts disclosure in a value chain by localizing reporting to an individual value chain participant among the plurality of value chain participants and direct suppliers, said first data inclusive of second data indicative of actual approval-request results of applications for regulatory approval and other data including assigned weights and weighted data, the RCS further operable to use a distributed ledger to ensure that responses from the plurality of regulator are securely recorded and reconstructable;
predicting with a RCAP (Regulatory Compliance Approval Prediction) module comprising a machine learning engine, predictive data including a probability that an application for regulatory approval by at least one regulatory agency of a value chain participant among said plurality of value chain participants will be approved by a regulator of the at least one regulatory agency among said plurality of regulators based on said first data collated and collected from said plurality of regulators and contained in said data structure including said second data indicative of actual approval-request results of applications for regulatory approval and the other data including the assigned weights and the weighted data, the predictive data including an outcome determination and a probability score indicative of a confidence of the predictive data for the at least one regulatory agency, wherein the machine learning engine further comprises a dimensional analysis and prediction engine that provides dimensional history and correlations and wherein analysis by the RCAP module is granular and the RCAP module communicates with the RCS; and
compiling the predictive data for improvements in predictive accuracy, wherein the compiling of the predictive data comprises assigning the weights to the machine learning engine based on an accuracy of at least one predictive classification algorithm in predicting an actual outcome with respect to the actual approval-request results of applications for regulatory approval.