A Novel Data Mining Testbed for User Centred Modelling and Personalisation of Digital Library Services

Abstract

Digital libraries can provide information services for users with diverse needs. Due to a large amount of data that exists in digital library systems, including text and multimedia resources, with different cohorts of users, and the challenges with existing digital library systems in terms of maintaining privacy and confidentiality, it is very difficult to provide personalised library services and improved user experience. However, novel data mining algorithms based on automatic user segmentation and borrowing behaviour modelling can leverage the relationship between users and borrowing records, to improve the library services. In this paper, we present an automatic approach for personalising the resources by segmenting the users and their preferences, based on a data mining strategy, involving, the classification based on Naïve Bayes, J48 and K-Nearest Neighbours Classification (K-NN) and using open source technology tools for evaluating the personalisation and improved user experience with digital library services.