Designing a new-type thesaurus-structure

Abstract

The paper aims at proposing a method to construct a structure combining the alphabetically arranged descriptors, supplied with individual identification marks, with a relatively large number of non-descriptors and all feasible permuted variations of the former so that the structure may adequately mirror the hierarchical (generic – specific), non-hierarchical (affinitive – preferential), as well as permutative relations of all concepts (terms) covered.
The author points out that this structure may be properly built up only after the hierarchical arrangement of the terms (hierarchical sub-system) by a graphical representation of the term-groups (graphical sub-system).
The author looks upon the whole of the thesaurus as a system, and considers both the alphabetical main form and the hierarchical resp. the graphic form necessary to compile a thesaurus suitable for serving several Information purposes.
The author also tackles the practical problems of constructing the thesaurus structure with a special view to the notations of terms, which may be handled by computer, too.
He also establishes a new relation-type, which he terms as permutative relation, and describes the characteristic features and role of this relation, as well as its merging with the main – alphabetical – part. The relation-type is deemed significant by the author both as regards the practical use of the thesaurus and its analysis.
All this is illustrated by means of an elaborated model containing a minor group of terms taken from labour management (it only contains 39 descriptors).The author also shows the algorithm of thesaurus construction and an elaborate graphic map of the model.
By analyzing the constructed model, he tends to determine certain essential thesaurus parameters (e.g. redundancy factor, level of term organization).
Using larger models, the author continues his investigations expecting to reach generalizable results.