A Memory Retrieval Aid for Hypothesis Generation.
Abstract
This study investigates an aid for the hypothesis retrieval process which is based on a model for hypothesis retrieval developed by Gettys, Fisher, and Mehle (1978). A computer simulates the human hypothesis retrieval process by searching an enriched associative memory which contains the associations of a number of individuals in the form of a lists of hypotheses for each datum. When the data of a decision problem become known, the appropriate lists are searched by the computer. Hypotheses that are common to most or all of the lists are suggested to the user, who assesses them for plausibility. An experiment was performed to determine the utility of the aid for both expert and non-expert users. The aid produced a substantial gain in performance for both groups of users, suggesting that further development of the aid would be worthwhile in decision situations which are repeated often enough to warrant the creation of an enhanced artificial memory. Also discussed are several techniques for implementing the aid, and determining the maximum gain in performance that the aid can produce.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 27, 1979
- Accession Number
- ADA072657
Entities
People
- Carol Manning
- Charles Gettys
- Stanley Fisher
- Suzanne Baca
- Tom Mehle
Organizations
- University of Oklahoma