ARHIVSKI VJESNIK 41. (ZAGREB, 1998.)

Strana - 69

S, Nikolovska, B. Asprovska, S. Dragićević, O. Iliev, The integrated information system of the Archives of Macedonia, Arh. vjesn., god. 41 (1998), str. 67-86 tic approach draws its experience form smaller organizations and through a scientifically supported generalization, it sets up the basics of modern archive work. This direction was later supported by numerous research work in this field (Roper - 72, Evans - 83, Posner - 76). They raise the level of archive work from mere preservation of documents and their use, mainly for historical research, towards a public service for current administration and citizens. In this way an integral approach was made possible towards the research of historical and current archive materials. The development of computers at the end of the sixties, and of the media for mass data storage and processing, gives the stimulus for the development of new conceptual models for efficient archive administration (Shellenberg - 75, Benedon - 79). Still, the implementation of computer systems becomes attractive in this field at the beginning of the nineties, as a result of: the sudden fall in prices of mass data storage media, of the new user-based operational systems, development of application programmes for multi-media data processing (processing of text, picture, sound, etc.). As one of the currently most relevant problems in the implementation of computer systems for archive work is the problem of data search. The problem arises from the impossibility to set up archive materials in an efficient structural framework (categorization) which would meet the needs of various kinds of historical research (Cook - 85), (Dojcinovski et al. - 96), the lack of precision in the presented data, overlapping, non-existence of a uniform terminology and search procedures. This problem is known as interdisciplinary searches (Weisberger - 93). The implementation of the concept of fuzzy sets (Chang - 92) and the cognitive sciences (Kuhlthau - 93) give a solid base for the development of efficient algorithms for interdisciplinary searches. At present, there are three different methods for solving this problem: - Method of weight vectors In this method, documents are represented as weight vectors in models based on the frequency of appearance (Forsyth & Rada - 86), (Salton - 70), binary probability models (Wong & Yao - 90), non-binary represented probability models (Meng & Park - 89), and fuzzy logic-based models (Salton - 83, Zadeh - 89). In agreement with this, searches will depend upon the value of the similarity function (metric distance) between the vector representations of the documents and the questionnaires. - Method of sentence collection This method uses a model of logic information models (Van Rijsbergen 87), where a premise is defined based on which the questionnaire can be derived. 69