Comparing Two Algorithms for Document Retrieval using Citation Links

This investigation compared the retrieval effectiveness of a one‐pass algorithm which employs a simple similarity calculation to that of the more complex multi‐pass Schiminovich algorithm. By means of lists of proposed relevant citations from users, searches on the AIP SPIN data base were conducted,...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Journal of the American Society for Information Science Ročník 28; číslo 4; s. 192 - 195
Hlavní autoři: Bichteler, Julie, Eaton, Edward A.
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York, N.Y American Documentation Institute 01.07.1977
Wiley Periodicals Inc
Témata:
Algorithms
Automatic classification
Automatic subject indexing
Bibliographic coupling
Citation analysis
Citations
Classification
Classification schemes
Clustering
Documents
Evaluation
Information storage
Information storage and retrieval
Information work
Mathematical techniques
Multi pass algorithms
One pass algorithms
Physicists
Physics
Retrieval
Schiminovich algorithm
Science and technology
Similarity
Single pass algorithms
Statistical techniques
Subject indexing
Technical services
ISSN:0002-8231, 1097-4571
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:This investigation compared the retrieval effectiveness of a one‐pass algorithm which employs a simple similarity calculation to that of the more complex multi‐pass Schiminovich algorithm. By means of lists of proposed relevant citations from users, searches on the AIP SPIN data base were conducted, resulting in two ordered lists of documents for each request. Recall/precision plots, based on user evaluations, illustrated that the average performance of the Schiminovich algorithm is marginally worse than that of the similarity algorithm. Thus, a relatively inexpensive procedure is sufficient to provide quite satisfactory document retrieval performance from citation data alone.
Bibliografie:ObjectType-Article-2
SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 14
ObjectType-Statistics/Data Report-1
ObjectType-Article-1
ObjectType-Feature-2
content type line 23
ISSN:0002-8231
1097-4571
DOI:10.1002/asi.1977.28.4.192