Intelligent Search in Digital Libraries

Abstract

The paper explores the role of an intelligent software agent as a mediator in information retrieval (IR). Software agents are autonomous programs which perform a specific personalized task in a heterogeneous and distributed environment. These autonomous, intelligent agents make decisions on behalf of the user, by narrowing the search domain and decreasing the human computer interaction, phenomenally. With the increasing number of institutes and universities across the country and globe it has become very important to share library services rather than duplicating the library facilities everywhere and incurring huge expenditure. In the present paper, the architecture for a digital library DIGLIB is proposed, which helps a user or a group of users identify and find reading material in a virtual or brick and mortar library of his or her interest. DIGLIB uses a software agent which is a unique combination of filtering and information agent to facilitate intelligent search. The information agent performs a general search on various sites rigorously based on a keyword and phrases. The short listed sites are further subjected to a specific search by the filtering agent. The information agent performs a breadth first search on the links ahead and also provides navigation recommendation. A rule-base is maintained by the filtering agent using premises like broad area of interest, specific area of interest, type of article, book, magazine, journals, periodicals etc. to be searched for further narrowing the search domain and suggesting URNs of digital libraries of user’s interest. An attempt has been made to design a digital library interface using an intelligent agent which continually runs in the background and facilitates the task of finding a suitable match between the desired library service and user’s requirement.