Generating Estimates by Combining Unsupervised and Supervised Machine Learning

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 31st, 2026

Detected March 31st, 2026

Summary

USPTO granted patent US12591802B2 to Intuit Inc. on March 31, 2026. The patent covers methods for generating estimates by combining unsupervised clustering and supervised machine learning, where peer entities are selected based on distance metrics and a trained model predicts target entity values. The patent names Jingxian Liao, Wei Wang, and Zhicheng Xue as inventors and includes 20 claims.

View original document View source feed page

What changed

USPTO granted patent US12591802B2 to Intuit Inc. covering methods that combine unsupervised and supervised machine learning for generating estimates. The method involves obtaining a cluster of reference entities, calculating distances between target entity features and reference entity features, selecting peer entities based on those distances, and applying a trained ML model to generate estimated metric values. The patent has 20 claims covering CPC classifications G06N 20/00, G06F 18/2148, and G06F 18/24137.

Companies developing similar machine learning methods for entity comparison or estimation should review the patent claims to assess potential design-around requirements or licensing needs. This is a routine patent grant with no compliance deadlines or regulatory obligations imposed on third parties.

Source document (simplified)

← USPTO Patent Grants

Generating estimates by combining unsupervised and supervised machine learning

Grant US12591802B2 Kind: B2 Mar 31, 2026

Assignee

Intuit Inc.

Inventors

Jingxian Liao, Wei Wang, Zhicheng Xue

Abstract

A method may include obtaining a cluster. The cluster may include a subset of reference entities. The method may further include calculating distances between features of a target entity and features of the subset of reference entities, selecting, based on the distances, peer entities from the subset, and generating an estimated value of a metric. The generating may include applying, to the features of the target entity, a machine learning model trained using training data including values of the features for the peer entities labeled with a value of the metric. The method may further include presenting the estimated value of the metric.