Discriminant forest classification method and system

US8306942B2 · kind B2 · utility

Assignee

• Lawrence Livermore National Security, LLC · US

Inventors

• Barry Chen · San Diego, US
• William G. Hanley · Livermore, US
• Tracy D. Lemmond · Tracy, US
• Lawrence J. Hiller · Livermore, US
• David A. Knapp · Livermore, US
• Marshall J. Mugge · Westerly, US

Key dates

• Filing date: May 6, 2009
• Grant date: Nov 6, 2012
• Expiry date: Mar 10, 2031

Classification

• Technology area (CPC G): Physics
• CPC primary: G06N20/20
• WIPO field: Computer technology
• WIPO sector: Electrical engineering

Abstract

A hybrid machine learning methodology and system for classification that combines classical random forest (RF) methodology with discriminant analysis (DA) techniques to provide enhanced classification capability. A DA technique which uses feature measurements of an object to predict its class membership, such as linear discriminant analysis (LDA) or Andersen-Bahadur linear discriminant technique (AB), is used to split the data at each node in each of its classification trees to train and grow the trees and the forest. When training is finished, a set of n DA-based decision trees of a discriminant forest is produced for use in predicting the classification of new samples of unknown class.