A Combined Stochastic and Deterministic Approach for Classification Using Generalized Mixture Densities.

Abstract

This work investigates a combined stochastic and deterministic optimization approach for multivariate mixture density estimation. Mixture probability density models are selected and optimized by combining the optimization characteristics of a multiagent stochastic optimization algorithm based on evolutionary programming and the expectation-maximization algorithm. Unlike the traditional finite mixture model, generally composed of a sum of normal component densities, the generalized mixture model is composed of shape-adaptive components. Rissanen's minimum description length criterion provides the selection mechanism for evaluating mixture model fitness. The classification problem is approached by optimizing a mixture density estimate for each class. A comparison of each class's posterior probability (Bayes rule) provides the classification decision procedure. A classification problem is posed, and the classification performance of the derived generalized mixture models is compared with the performance of mixture models generated using normally distributed components. While both approaches produced excellent classification results, the generalized mixture approach produced more parsimonious density models from the training data. (KAR) P. 1

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 1995
Accession Number
ADA296703

Entities

People

  • D. E. Waagen
  • J. R. Mcdonnell

Organizations

  • Naval Command, Control and Ocean Surveillance Center

Tags

Communities of Interest

  • Human Systems
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Artificial Intelligence
  • Computer Programming
  • Estimators
  • Information Science
  • Kernel Functions
  • Machine Learning
  • Mathematics
  • Neural Networks
  • Operations Research
  • Optimization
  • Probability
  • Probability Density Functions
  • Probability Distributions
  • Random Variables
  • Statistics
  • Training

Fields of Study

  • Engineering

Readers

  • Operations Research
  • Statistical inference.