A comparative evaluation of four thesauri concerning “conceptual representability” was conducted. These thesuari were ERIC Thesaurus, JICST Thesaurus and two originally constructed thesauri of library and information science. For this purpose, the following issues come under review first: the evolution of thesauri, the history of international guidelines for thesaurus establishment, the functions of thesauri, the new trend for various thesauri as switching languages or integrated vocaburalies, and methods for evaluation of thesauri. These lead us to regard thesauri as what represent the knowledge structures of the fields. For evaluating thesauri in this meaning, “conceptual representability”, which indicates how appropriately the thesaurus can represent the concepts in documents or in queries, is one of the important propeties of thesauri. It has two aspects, “comprehensivity” and “specificity”.
We have conducted an experiment in this aspect as follows: The sample documents were extracted randomly from the fields of library and information science. The sample documents were indexed according to the four thesauri mentioned above. The list of indexing terms were compared and “conceptual representability”, “comprehensivity” and “specificity” were computed. Through classifying the concepts appeared in the document we have got several subject categories. It was clarified that these three indeces, “conceptual representability”, “comprehensivity” and “specificity”, revealed the advantages and disadvantages of each thesauri, and that the number of terms involved in the thesauri is proportional to the “conceptual representability”. This indicates that definite number of terms, discussing here, is necessary for vocaburalies in order to represent the concepts appeared in a field.
© 1989 三田図書館・情報学会© 1989 Mita Society for Library and Information Science
This page was created on 2022-03-16T16:25:14.306+09:00
This page was last modified on