A Naive Bayes-Based Model for Predicting Users' Information Needs in Digital Library Systems

The contemporary digital library systems have been functioning in the environment that is characterized by the rapidly increasing volume of information resources, so it is becoming more challenging to be able to offer the users the relevant materials in the most efficient way. Classical methods of search only react to explicit user requests and do not always detect the hidden or new information requirements. This paper represents a smart model to predict the information needs of the users in library systems based on the Naive Bayes classification algorithm. The proposed model examines demographic, academic and behavioral user characteristics such as age, profession, education, search history, download and even preferred subject areas to predict the likelihood of user interest in a particular type of information. The system will create individualized recommendations of resources based on the projected information requirements. The experimental analysis of the approximate accuracy of the proposed method based on user data in the Libsmart digital library system indicates that the given method has an accuracy of 86.4, which proves its efficiency in providing adaptive and user-based information services. The intelligent recommendation modules in digital library setting based on the proposed model can act as a viable basis.

Ҳали таржима қилинмаган