Download PDF
Methods Inf Med 1987; 26(01): 31-39
DOI: 10.1055/s-0038-1635480
Original Article
Schattauer GmbH

A Method of Medical Knowledge Base Augmentation

Eine Methode zur Erweiterung medizinischer Wissenssammlungen
S. Lester
1   From the School of Medicine, University of Washington, Seattle, and the National Library of Medicine, Bethesda, MD, USA
,
R. Rada
1   From the School of Medicine, University of Washington, Seattle, and the National Library of Medicine, Bethesda, MD, USA
› Author Affiliations
Further Information

Publication History

Publication Date:
16 February 2018 (online)

    Permissions and Reprints

Summary

The Medical Subject Headings (MeSH) of the National Library of Medicine may be viewed as a semantic network. The relationships in this semantic network are of a broader-than/narrower-than type. A knowledge base of this type may be augmented by adding new terms and new relationships to the network. The Current Medical Information and Terminology (CMIT) of the American Medical Association represents a rich source of relationships for the disease terms of MeSH. A subset of MeSH was augmented with the knowledge from a subset of CMIT using a matching and similarity strategy. The matching portion of the experiment showed that about half of CMIT may be directly merged with MeSH based on exact and partial matches and utilization of alternate and synonym terms from CMIT. The similarity portion of the experiment showed that a method of merging based on similarity of features is a workable approach to incorporating knowledge into MeSH when lexical matches are not available. Evaluation of the resulting merged knowledge base suggested that the etiology property of CMIT was the most highly inherited property. The augmented knowledge base was used as a basis for an automatic indexer. The indexer was less accurate after augmentation than before. One key difficulty stemmed from the way that CMIT was encoded into MeSH. More powerful encodings of CMIT into MeSH are being pursued. Building on MeSH, CMIT, and other such knowledge bases that already exist on the computer is one way to try to develop intelligent medical computer systems.

Die Sammlung von medizinischen Schlagwörtern (Medical Subject Headings = MeSH) der US-Staatsbibliothek für Medizin kann als ein semantisches Netzwerk betrachtet werden. Die Relationen in diesem semantischen Netzwerk sind von der Art »weiter als« und »näher als«. Eine Wissensdatenbasis dieser Art kann durch Hinzufügen neuer Begriffe und Relationen vergrößert werden. Die »aktuelle medizinische Information und Terminologie« (CMIT) der Amerikanischen Medizinischen Gesellschaft (AMA) ist eine reiche Quelle für Relationen zwischen den Krankheitsbegriffen von MeSH.

Eine Untergruppe von MeSH wurde mit den Informationen einer Untergruppe von CMIT unter Verwendung einer Matching- und Gleichheitsstrategie vergrößert. Der erste Teil dieses Versuchs zeigt, daß etwa die Hälfte von CMIT direkt mit MeSH verbunden werden kann, entweder durch eine exakte oder eine teilweise Anpassung oder durch den Gebrauch von alternierenden und synonymen Begriffen von CMIT. Die Gleichheitsstrategie in diesem Versuch zeigt, daß eine Methode des Verbindens, die auf Gleichheit von Kennzeichen beruht, ein Ansatz ist, um Wissen in MeSH einzubauen, falls keine lexikalische Zuordnung zur Verfügung steht. Die Auswertung der erzeugten Datenbasis erweist, daß der Ätiologie-Anteil von CMIT der am besten übertragbare ist.

Die vergrößerte Datenbasis wurde als Basis für eine automatische Indizierung benutzt. Diese Indizierung war weniger gut als die, die mit der ursprünglichen Datenbasis durchgeführt wurde. Eine Hauptschwierigkeit resultierte aus der Art, wie CMIT in MeSH eingebaut wurde. Eine wichtige Aufgabe ist es, diesen Vorgang zu verbessern. Die Implementierung von MeSH, CMIT und anderen existierenden Wissensbasen in den Rechner ist ein Weg, intelligente medizinische Computersysteme zu entwickeln.

Key-Words

Machine Learning - Knowledge Bases - Semantic Nets - Automatic Indexing - Natural Language Processing

Schlüssel-Wörter

Lernende Maschinen - Wissensbasen - semantische Netze - automatisches Indizieren - Verarbeitung natürlicher Sprache
 

  • REFERENCES

  • 1 Blair D, Maron M. E. An Evaluation of Retrieval Effectiveness for a Full-Text Document-Retrieval System. Comm. Ass. Comput. Machin 1985; 28: 289-299.
    PubMedGoogle Scholar
  • 2 Blois M. S, Tuttle M. S, Sherertz D. D. Reconsider: A Program for Generating Differential Diagnoses. Proceed. Fifth Ann. Symp. Comput. Appl. med. Care (IEEE Computer Society). 1981
    Google Scholar
  • 3 Blois M. S. Information and Medicine: The Nature of Medical Descriptions. Berkeley, California: University of California Press; 1984
    Google Scholar
  • 4 Brachmann R. What IS-A Is and Isn’t: An Analysis of Taxonomic Links in Semantic Networks. Computer 1983; 16: 30-36.
    PubMedGoogle Scholar
  • 5 Buchanan B, Shortliffe E. Rule-Based Expert Systems: The MYCIN Experiments of the Stanford Heuristic Programming Project. Reading, MA: Addison-Wesley; 1984
    Google Scholar
  • 6 Buchanan B, Fu Li-Min. Learning Intermediate Concepts in Constructing a Hierarchical Knowledge Base. Proc. Ninth Intern. Joint Conf. Art. Intell. 1985: 659-666.
    Google Scholar
  • 7 Clancey W, Shortliffe E. Introduction: Medical Artificial Intelligence Programs. In Shortliffe E. (Edit.) Readings in Medical Artificial Intelligence. Reading, MA: Addison-Wesley; 1984: 1-18.
    Google Scholar
  • 8 Eng J, Lu F, Rada R, Brylawski B. Morpliosemantic Analysis in Indexing. MEDINFO-86.. Amsterdam: North Hol-land; 1986
    Google Scholar
  • 9 Fikes R, Kehler T. The Role of Frame-Based Representation in Reasoning. Comm. ACM 1985; 28: 904-920.
    CrossrefPubMedGoogle Scholar
  • 10 Finkel A, Gordon B, Baker M, Fanta C. Current Medical Information and Terminology. Chicago: American Medical Association; 1981
    Google Scholar
  • 11 Forsyth R, Rada R. Machine Learning: Expert Systems and Information Retrieval. London: Ellis Horwood; 1986
    Google Scholar
  • 12 Lancaster F. W. Vocabulary Control for Information Retrieval. Washington, D. C: Information Resources Press; 1972
    Google Scholar
  • 13 Luhn H. P. The Automatic Creation of Literature Abstracts. IBM J. Res. Dev 1958; 2: 159-165.
    CrossrefPubMedGoogle Scholar
  • 14 McCarn D. B. MEDLINE: an Introduction to On-Line Searching. J. Amer. Soc. Inform. Sci 1980; 31: 181-192.
    CrossrefPubMedGoogle Scholar
  • 15 Medical Subject Headings Section of National Library of Medicine. Medical Subject Headings. Tree Structures 1986 NLM-MED-85-03 (Bethesda, MD, NLM). 1985
    PubMedGoogle Scholar
  • 16 Mili H, Rada R. A Statistically Built Knowledge Base. Proceed. Expert Systems Gov. Conf. IEEE Computer Society Press; 1985: 457-463.
    Google Scholar
  • 17 Miller P. L. A Critiquing Approach to Expert Computer Advice: ATTENDING. London: Pitman Publishing; 1984
    Google Scholar
  • 18 National Library of Medicine. MEDLARS, the Computerized Literature Retrieval Service. DHEW NIH. Bethesda, MD: U.S. Gov. Print. Off; 1979: 79-1286.
    Google Scholar
  • 19 Nilsson N. Principles of Artificial Intelligence. Palo Alto, CA: Tioga; 1980
    Google Scholar
  • 20 Politikas P, Weiss S. Using Empirical Analysis to Refine Expert System Knowledge Bases. Artif. Intell 1984; 22: 23-48.
    CrossrefPubMedGoogle Scholar
  • 21 Quillian M. R. Semantic Memory. In Minsky M. (Edit.) Semantic Information Processing.. Cambridge, MA: MIT Press; 1968
    Google Scholar
  • 22 Rada R, Evans L. Automated Problem Encoding System for Ambulatory Care. Comput. biomed. Res 1979; 12: 131-139.
    CrossrefPubMedGoogle Scholar
  • 23 Rada R, Humphrey S, Coccia C. A Knowledge-base for Retrieval Evaluation. Proc. ACM ‘85. 1985: 360-367.
    Google Scholar
  • 24 Rada R, Darden L, Eng J. Relating Two Knowledge Bases: The Role of Identity and Part-Whole. Proceed. Sec. Ann. Symp. Role of Language in Problem Solving. (forthcoming).
    PubMedGoogle Scholar
  • 25 Rada R. Gradualness Eases Refinement of Medical Knowledge. Med. Inform 1986; 11: 59-73.
    PubMedGoogle Scholar
  • 26 Rescher N. Axioms of the Part Relation. Philosoph. Stud 1955; 6: 8-11.
    PubMedGoogle Scholar
  • 27 Sager N. Natural Language Information Processing. Reading, MA: Addison-Wesley Publishing; 1981
    Google Scholar
  • 28 Salton G, McGill M. Introduction to Modern Information Retrieval. New York: McGraw-Hill; 1983
    Google Scholar
  • 29 Schank R, Kolodner J, DeJong G. Conceptual Information Retrieval. In Williams P. W. (Edit.) Information Retrieval Research. London: Butter-worths; 1981: 94-116.
    Google Scholar
  • 30 Soergel D. Indexing Languages and Thesauri: Construction and Maintenance. New York: Wiley; 1974
    Google Scholar
  • 31 Sridharan N. S. Evolving Systems of Knowledge. AI Magazine 1985; 108-119.
    PubMedGoogle Scholar
  • 32 Stassen H. H. The Similarity Approach to EEG Analysis. Meth. Inform. Med 1985; 24: 200-212.
    PubMedGoogle Scholar
  • 33 Touretzky D. The Mathematics of Inheritance Systems. London: Pitman Publishing; 1986. also appeared as Ph. D. thesis from Carnegie-Mellon University, Dept. Computer Science,1984
    Google Scholar
  • 34 Wessel A. Computer-Aided Information Retrieval. New York: 1975
    Google Scholar
  • 35 Wingert F. An Indexing System for SNOMED. Meth. Inform. Med 1986; 25: 22-30.
    PubMedGoogle Scholar
  • 36 Winston P, Horn B. LISP. Reading, MA: Addison-Wesley; 1984
    Google Scholar
  • 37 Winston P. Artificial Intelligence. Reading, MA: Addison-Wesley; 1984
    Google Scholar