The ROC Curves of Fused Independent Classification Systems

Abstract

The need for optimal target detection arises in many different fields. Due to the complexity of many targets, it is thought that the combination of multiple classification systems, which can be tuned to several individual target attributes or features, might lead to more optimal target detection performance. The ROC curves of fused independent two-label classification systems may be generated by the mathematical combination of their ROC curves to achieve optimal classifier performance without the need to test every Boolean combination. The monotonic combination of two-label independent classification systems which assign labels to the same types results in a lattice of ROC curves which are epimorphic to the corresponding combinations of classification systems. Provided the ROC curves of individual systems are available, testing the lattice of ROC curves in software with existing individual ROC curves can represent a significant cost savings in the design of optimal classification systems.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 2008
Accession Number
ADA486884

Entities

People

  • Michael B. Walsh

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • C4I
  • Energy and Power Technologies
  • Materials and Manufacturing Processes
  • Sensors
  • Space

DTIC Thesaurus Topics

  • Air Force
  • Algorithms
  • Applied Mathematics
  • Boolean Algebra
  • Classification
  • Data Fusion
  • Data Sets
  • Detection
  • Detectors
  • Identities
  • Machine Learning
  • Mathematics
  • Numbers
  • Probability
  • Real Numbers
  • Target Detection
  • Trajectories

Fields of Study

  • Computer science

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Computer Vision.
  • Regression Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Neural Networks