Dalrymple, P.W.; Cox, R.: ¬An examination of the effects of non-Boolean enhancements to an information retrieval system (1992)
0.00
0.0036430482 = product of:
0.014572193 = sum of:
0.014572193 = weight(_text_:information in 2939) [ClassicSimilarity], result of:
0.014572193 = score(doc=2939,freq=12.0), product of:
0.06134496 = queryWeight, product of:
1.7554779 = idf(docFreq=20772, maxDocs=44218)
0.034944877 = queryNorm
0.23754507 = fieldWeight in 2939, product of:
3.4641016 = tf(freq=12.0), with freq of:
12.0 = termFreq=12.0
1.7554779 = idf(docFreq=20772, maxDocs=44218)
0.0390625 = fieldNorm(doc=2939)
0.25 = coord(1/4)
- Abstract
- One of the problems in information retrieval (IR) research is that few of the non-Boolean features of experimental systems developed by IR researchers have been adopted in commercially available systems which can be evaluated using real users with actual information needs. Without the opportunity to examnine how these features perform with actual bibliographic files and how they affect users in their information-seeking tasks, our understanding of information retrieval remains limited, and system development fails to advance. The research describes here compared two CD-ROM MEDLINE systems for the Macintosh, one of which incorporates many of the features previously identified by research as central to sound and innovative IR design such as: elimination of the need for Boolean logical connectors, acceptance of a natural language query, and ranked output. The other is more traditional in its design. Two groups of search topics selected from the National Library of Medicine's test queries in clinical medicine were searched using both a natural language strategy and a strategy based on MeSH vocabulary. results were compared on the following variables: search input and processing times, set size, overlap between sets produced by the two systems, and evaluative judgements made by subject experts. The findings indicate the these systems differ on these dimensions, and greater variance occurs in the natural language searches
- Imprint
- New York : Learned Information
