SEQUENTIAL DESIGNS OF EXPERIMENTS FOR ALTERN- ATIVE OBJECTIVE FUNCTIONS IN AUTOMATED TEACHING PROGRAMS

Abstract

Presents the theoretical foundation, in terms of statistical decision theory, for the development of branching procedures or designs of automated teaching programs best tailored to individual student needs. Attacks the design problem for automated teaching programs or experiments from the standpoint of the theory of the sequential design of experiments. Outlines the general theory of the sequential design of experiments and the use of Bayesian procedures for determining best designs. Outlines the technique of solution for best sequential designs of experiments called ''backward induction''. Discusses the characteristics that models of teaching processes need to have in order to be accessible to computation for best designs in full-scale teaching programs even when the back ward induction technique is applied. Emphasizes the critical importance of coarse sufficient partitions of the sample space of teaching models.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 1963
Accession Number
AD0408549

Entities

People

  • Robert E. Dear

Organizations

  • System Development Corporation

Tags

Communities of Interest

  • C4I
  • Human Systems

DTIC Thesaurus Topics

  • Computational Processes
  • Computations
  • Computer Programs
  • Decision Theory
  • Experimental Design
  • Game Theory
  • Markov Chains
  • Mathematical Models
  • Models
  • Probability
  • Probability Distributions
  • Random Variables
  • Sequences
  • Sequential Games
  • Statistics
  • Stochastic Processes
  • Zero-Sum Games

Readers

  • Artificial Intelligence
  • Regression Analysis.
  • Software Engineering

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • Space