We identified general indexing concepts, based on descriptors (preferred-terms) which contain terms 'agriculture' and 'agricultural', in several non-agricultural and general information systems/databases, and respective thesauri (controlled vocabularies), covering fields of civil, mechanical, chemical engineering, physics, psychology, medicine, biomedicine, education, business, economics, finance, library science, sociology, social and political sciences and related disciplines (for example, CSA-Illumina ERIC (Education Resources Information Center), LISA (Library and Information Science Abstracts), Sociological Abstracts (Sociological Indexing Terms), Ebsco Academic Search Complete (Subject Terms), Medline (MeSH), Political Science Complete, Ei Engineering Village Compendex and Inspec, PsycINFO (Thesaurus of Psychological Index Terms), ABI/Inform Global (ProQuest Thesaurus).
We compared characteristics, strengths and limitations of thesauri in each respective system with regard to searching (word-, phrase-indexing, truncation (wildcard), stemming, autostemming, hierarchical structure (tree-structures) and relations among preferred-, non-preferred terms, Narrower, Broader and Related Terms. We assessed database coverage of general agriculture-related topics based on retrieval with these terms.
A very interesting paper about the use of different search terms and phrases in different collections. Very informative for a non-librarian, but also a bit difficult to understand sometimes. It is also very relevant to learn that agricultural subjects are found in many collections, which might otherwise be considered irrelevant.
ReplyDeletesuggestions for changes: in fig. 1 and 2, it is very difficult to tell which "color" is which (they are all grey). They should be made more distinct. Also in both figures, there is no label on the x-axis. In fig. 1, it is easy to guess that the label is pretty much the same as the fig. text, but it is not so easy in fig. 2. In the text for fig. 1, (BT) is shown twice, one of them should probably be (RT), and on the figure it says NT1 and BT1 - why?
I find this paper really interesting, as it provides detailed information on various thesauri used in the agricultural sector. The work carried out and presented in this paper could prove useful for using it as a basis for a more comprehend work on the specif area.
ReplyDeleteIt looks really relevant to the scope of the VOA3R project as well!
Since Thomas started by mentioning that young people only search Google (not even Google Scholar), it would be interesting if the results from the different thesauri was compared with the outcome of a search in Google. While this question was answered at the session, it would be really good to include in the introduction and discussion of the paper.
ReplyDeleteIlse, thank you for your comments. Considering your suggestions for changes in the fig. 1 and 2 I have realized that there must have been an apparent technical error during the conversion of my original pdf file for the purposes of this blogspot. My originally sent manuscript, which has also been published in the Proceedings, shows distinct colors and patterns for all items . That 'related terms (BT)' should indeed read 'related terms (RT)'. A typing error. NT1 and BT1 are usual abbreviations which denote the narrower and broader terms on the first lower/upper level. I only counted the first level terms because there are more terms on further levels of the hierarchical tree.
ReplyDeleteAn interesting paper which shows the importance of using many different databases in order to retrieve comprehensive information on agriculture. Also it comprehensively explains that indexing systems are quite different in different database, and its implication for retrieval
ReplyDelete