Abstract
In this paper, we introduce ACS: an automatic classifica tion system for school libraries. First, various approaches towards automatic classification, namely (i) rule-based, (ii) browse and search, and (iii) partial match, are critically reviewed. The central issues of scheme selection, text analysis and similarity measures are discussed. A novel approach towards detecting book-class similarity with Modified Overlap Coefficient (MOC) is also proposed. Finally, the design and implementation of ACS is presented. The test result of over 80% correctness in automatic classi fication and a cost reduction of 75% compared to manual classification suggest that ACS is highly adoptable.
Get full access to this article
View all access options for this article.
