LEARNING WITH A LACK OF PRIOR DATA

Abstract

A convenient model for learning is provided by the sequential compound decision problem of mathematical statistics. The decision-maker observes a sequence of independent random variables, the distribution of which varies arbitrarily along the sequence. Since the decision-maker does not know the distribution beforehand, he tries to learn during the sequence how to minimize his losses.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1967
Accession Number
AD0828011

Entities

People

  • Bruno O. Shubert

Organizations

  • Stanford University

Tags

DTIC Thesaurus Topics

  • Air Force
  • California
  • Contracts
  • Convergence
  • Distribution Functions
  • Electronics
  • Electronics Laboratories
  • Inequalities
  • Learning
  • Numbers
  • Probability
  • Random Variables
  • Real Numbers
  • Security
  • Sequences
  • Universities

Fields of Study

  • Mathematics

Readers

  • Neural Network Machine Learning.
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.
  • Theoretical Analysis.