A new type of information retrieval system

  • Authors:
  • H. Joel Jeffrey

  • Affiliations:
  • Vanderbilt University, Nashville, Tennessee

  • Venue:
  • ACM-SE 14 Proceedings of the 14th annual Southeast regional conference
  • Year:
  • 1976

Quantified Score

Hi-index 0.00

Visualization

Abstract

In the period 1964-1968, Peter G. Ossorio [10,11,12] developed and tested, on a pilot study basis, a new approach to the problem of automatic document retrieval. Ossorio's studies were entirely successful, as pilot studies, and show the feasibility of using his approach to produce a new kind of retrieval system.These retrieval systems do not operate by word matching. The basic approach is to simulate the judgement of competent human judges of the conceptual content of each document, and the request. This judgement is then used to retrieve those documents with conceptual content most similar to that of the request.Each document is processed only at the time it is added to the data base, in time linear in the number of words in the document that the system recognizes. The retrieval request is in ordinary English. Time for retrieval is linear in the number of documents on file. Documents are retrieved in order of similarity of conceptual content to that of the request. The system works, in certain respects, better on full text documents, providing better descriptions of document content, and more detailed cross-indexing.The new type of system shows a number of interesting features. Among these are:(1) Much better performance than systems using the old techniques;(2) Faithful representation of the judgement of the person(s) whose judgement is being simulated, thus providing the possibility of individualized retrieval systems;(3) Ability to explain to a user why it retrieved certain documents, and not others. With this information, the user can alter his request, or instruct the system to judge things differently;(4) Automatic recognition of requests the system cannot properly handle;(5) Sub-documentary indexing reflecting heterogeneity of material. As is often the case with a new paradigm, Ossorio's work raises at least as many questions as it answers. This paper presents the new approach, and the results of some first explorations in the new field.