DYNAMIC INFERENCE.

Abstract

When from the pattern of observable variables the probability distributions of process parameters is inferred and then these distributions are used t pedict the future course of the process; this is called Dynamic Inference. For a military illustration of this problem the case of detecting subbmarines using a fixed sonar station is considered. A business application can arise in marketing. The problems discussed are illustrated graphically. Two stochastic processes are pictured, the first providing the parameters of the second, and the second having an observable output. There are some problems where the parameters previously generated by the first stochastic process affect the future parameters generated by the process, and others where previous observable variables can affect future observable variables. Cases are possible where previous observable variables have an effect on the parameter generation process. All of these possiblilites are indicated by the delay pates. The model developed can theoretically treat all of these cases, but its present application is restricted to the case where none of the delay branches is present. (Author)

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1964
Accession Number
AD0458367

Entities

People

  • Ronald A. Howard

Organizations

  • Massachusetts Institute of Technology

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Commerce
  • Demographic Cohorts
  • Marketing
  • Mathematics
  • Probability
  • Probability Distributions
  • Stochastic Processes

Fields of Study

  • Mathematics

Readers

  • Control Systems Engineering.
  • Regression Analysis.
  • Theoretical Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms