Abstracts

The problem of collection, organization and retrieval of information within the framework of particular fields of investigation (ecology, botany, zoology, forestry, genetics, radiobiology, etc.) is one of the key problems among the questions of integrating resources in disciplines of natural sciences. Here two tasks arise: 1) development of certain clearly and unambiguously interpreted conceptual base, i.e. thesauri for suitable disciplines and 2) representation of information of cross-disciplinary character, falling outside the scope of a particular field, i.e. creation of retrieval systems with the most coverage of available materials and approaching to expert-reviewer for their "intelligence" potentialities. In the paper we shall dwell primarily on the solution of the second task though lately we gained certain experience in creation of electronic dictionaries.

Since 1991 an information bank on Chernobyl problems resulting from direct review and abstracting of literature published in Belarus, Russia and Ukraine is being accumulated at the Institute of Genetics and Cytology of the National Academy of Sciences of Belarus. The bank became the basis for developing database of cross-disciplinary information CHEDIBASE (CHErnobyl Digest Information BASE) and interdisciplinary bulletin "Chernobyl Digest" in Russian and English.

Since 1991 6 issues of "Chernobyl Digest" were prepared. This interdisciplinary information is of great interest for specialists – catastrophologists, biologists, ecologists, physicians, psychologists, etc. A need for such information and its effective delivery to users does not decrease with the time and makes to think of using resources of the most advanced information technologies for meeting reader demands at the up-to-date level. That is why since May 2000 the bulletin is arranged on BIOBEL, server of the Department of Biological Sciences at the National Academy of Sciences of Belarus (http://biobel.bas-net.by/igc/ChD).

The bulletin contains only a part of information accumulated in CHEDIBASE (from 30 to 50% of the available more than 4 thousand abstracts, depending on the language of the edition). Therefore, the problem is raised to make a remote access to the whole file of available information. As at the initial period we have used database management system (DBMS) FoxPro for CHEDIBASE formation, in 1998–1999 we transferred information to DBMS Access, more advanced and convenient in on line use.

At present the CHEDIBASE structure comprises all the data required for retrieval and systematization of information for both the content of work and auxiliary data (year of publication, source, page, number, etc.). It also has additional fields provided for carrying out content analysis of information. We have developed the system of descriptor code (DC) used for filling in particularly provided database fields that enable determination for every record of such parameters as object of investigation, subject of investigation, discipline where investigations were carried out and methods of investigation. Codes are assigned to information by expert-reviewer during its analytical processing and abstracting, and therefore each assigned descriptor is capable of implementing rather deep meaning retrieval. The results have shown that the DC meaning component in combination with retrieval-systematization potentialities of the applied DBMS provides the possibility of automated analytical selection of the data answering the required attribute or their combination. Till recently the possibility of substituting human in such an intelligence retrieval was called in question.

Distinctions and advantages of retrieval-sorting system using DC should be noted. One of the most commonly used methods for retrieval and sorting of information in modern databases is key word retrieval. This method is used in the majority of modern systems working with information files. As a rule, all these systems use any term from the most often occurred in particular work in the role of key word. DC sorting gives an opportunity to reduce the probability of blank in work possible at sorting by the above-mentioned method and taking place due to inadequacy terms applied by user and reviewer. The attribute, interesting the user and mainly described by a large synonymous set of words, will be defined only by one DC in the system proposed by us. Besides, the reference DC table, which is applied by both reviewer and user, is small by size (in comparison with required comprehensive directory of key words) and comfortably arranged.

The so-called “electronic encyclopaedias”, i.e. computer programs which allow storage of large quantity of documents (glosses) and their manipulation identifying texts according to available rubricator, are being developed and widely popularized at present. It should be noted that it is rather expensive product and usually one team is engaged in developing any encyclopaedia and another team transforms it into electronic form. People without humanitarian education have to go deep into humanitarian problems. Descriptor tables, used in CHEDIBASE, are in complete agreement with rubricator of electronic encyclopaedia but creation of a DC table can be done easier. Any specialist (i.e. biologist or geneticist) can create independently a descriptor table and then database without necessity to organize joint work with a team of mathematicians or programmers. It is enough to master basic procedures of work in DBMS Access, which is available in any computer system with Microsoft Office package installed, and major principles of compiling descriptor tables.

Full Text in Russian:

HTML

Note. Abstracts are published in author's edition

Comments Comments

[Home] [Conferences] [SB RAS]

© 2002, Siberian Branch of Russian Academy of Science, Novosibirsk
© 2002, UII SB RAS, Novosibirsk