Time Series Forecasting with Outlier Filtering and Model Selection

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 31st, 2026

Detected March 31st, 2026

Summary

USPTO granted US Patent 12591787B1 to United Services Automobile Association (USAA) on March 31, 2026. The patent covers an event forecasting system that filters time series data by identifying and handling outliers, selects predictive models through a two-stage ranking validation process, and provides a user interface for manual adjustment of predictions. The patent contains 20 claims under CPC classifications including G06N 20/00.

View original document View source feed page

What changed

The USPTO granted US Patent 12591787B1 to United Services Automobile Association (USAA) on March 31, 2026. The patent describes a time series forecasting system that filters data by identifying outliers aligned with exceptions, removes or replaces them, and selects optimal predictive models through a two-stage ranking process using different parts of the time series data for validation. The system also includes a user interface allowing users to view known and predicted values and manually adjust predictions. The patent, filed February 10, 2022 under Application No. 17668631, contains 20 claims and falls under CPC classifications G06N 20/00, G06F 30/20, G06F 18/214, and G06F 17/18.

This is a routine patent grant announcement rather than a regulatory action. Companies developing machine learning-based forecasting systems should review the patent scope to assess potential licensing considerations or design-around needs. No compliance deadlines or penalties are associated with this intellectual property grant. Organizations utilizing similar time series forecasting technology may benefit from conducting an IP landscape review.

Source document (simplified)

← USPTO Patent Grants

Forecasting events by modeling time series data

Grant US12591787B1 Kind: B1 Mar 31, 2026

Assignee

United Services Automobile Association (USAA)

Inventors

Austin Jenkins, Matthew Christopher Kaplan, Conor Wroble

Abstract

An event forecasting system can filer time series data by identifying outliers and/or data instances that align with an exception, and either remove or replace them. The event forecasting system can further select a set of models expected to be best predictive for the time series data by performing an initial ranking using a first part of the time series data, validated by a second part; and performing a second ranking of the best models using the first part and second parts of the time series data, validated by a third part. Finally, the event forecasting system can output results to a user interface where a user can view the known and predicted data values and make injects to manually adjust the predicted values as desired.