Content-Based Document Recommender System for Aerospace Grey Literature: System Design

Abstract

Recorded knowledge in the form of manuscripts, print documents, microforms, CD-ROMs, computer files, etc., is increasing exponentially. In order to locate and access relevant information from this vast amount of literature, efforts have been made from time and on to develop various tools and techniques. Early evolved tools/techniques include: library catalogues, indexes, concordances and so on. Information Retrieval Systems played a vital role in the field of information & librarianship to find relevant information from vast number of documents. Recently intelligent agents gained importance as they are able to query databases and resources on Internet, remote library catalogs thereby reducing information overload on the user. Another technology that alleviates the information overload problem is the filtering or recommender systems. The purpose of development of recommender systems is to provide useful and most relevant recommendations or suggestions from number of available alternatives. The present study aimed at design & development of Content-based Document Recommender System (CODORS) to retrieve most relevant technical documents without necessarily matching title terms and closely related to a particular search term(s) as opposed to general Online Public Access Catalog (OPAC) search results. The developed CODORS converts terms expressed by the user in natural language automatically into subject descriptors, carry on search, ranks and retrieves documents.

http://dx.doi.org/10.14429/djlit.31.3.1046