PROPERTY PREDICTION USING RESIDUAL STEPWISE REGRESSION

Patent №

US 6,804,609

Granted

2004-10-12

Filed 2003

Owner

CONOCOPHILLIPS COMPANY

Lab

AI components

3

ml · planning · evo

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

10412995

A method for predicting reservoir properties at an unexplored subterranean location based on known seismic attributes at the unexplored subterranean location, as well as known reservoir properties and seismic attributes at a plurality of explored subterranean locations proximate to the unexplored subterranean locations. The method uses residual stepwise regression to generate a prediction equation capable of calculating predicted reservoir properties based on multiple seismic attributes. The residual stepwise regression involves using residual values from a previous regression step to determine which seismic attribute will be included in a subsequent regression step.

AI classification

Planning1.00
Machine learning1.00
Evolutionary computation0.99
AI hardware0.15
Vision0.06
Knowledge representation0.01
Speech0.01
Natural language0.00

Ownership

CONOCOPHILLIPS COMPANY

assignment · 139770956

Assignors

BRUMBAUGH, DAVID L.

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC