Document subject matrix as a factor of precision in computerized information systems
Devices meant to improve IR systems precision have not paid much attention to the subject structure of the documents recorded in the databases, rather they have focused on indexing systems and retrieval strategies. It has been found that the subject matrix of a document - the pattern of relationships existing between the array of the subjects used as its descriptors - determines the chances of the document being relevant to the question for which it is retrieved. The subject matrices of most documents are such that interference of terms which result in fallouts would occur if coordination is done without recourse to the relationships existing between its descriptors as is the case when Boolean 'AND' is used in retrieval of when random pre-coordination takes place. The more complex the matrix, the higher the chances of the document being found irrelevant in an IR system output. The interference of incompatible subjects within a pool of descriptors of document record at the point of coordination is a major cause of low precision of IR systems. Improvement can be made in the performance of IR systems in terms of precision by increasing the intellectual input to indexing, especially to take into consideration the subject matrix of the document. Multiple entries should be made for any document that has a matrix of subjects that would cause it to be retrieved for questions for which it is irrelevant.
| Main Author: | |
|---|---|
| Format: | Journal Article biblioteca |
| Language: | English |
| Published: |
1990
|
| Subjects: | documentation, data processing, information retrieval, |
| Online Access: | https://hdl.handle.net/10568/101861 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|