A survey on automatic recommenders in the context of institutional repositories

Abstract

The present article provides a general review of automatic recommendations tools in digital libraries. In this context, we aim to analyse three elements: how digital libraries are conceptualized (that is, how they are represented as software entities), the algorithms used by automatic recommenders to perform their logic, and the modelling of users' interactions in this kind of environments. In order to define the complete context of the research, we will start by defining what a digital library is and the way it is represented in common software management applications. Then, we will review the basic definitions related to recommenders and the usual algorithms. We focus on hybrid recommendation techniques because they provide more accurate results and help to get over the inherent weak points of basic techniques used individually. Finally we will analyse the users' interactions in digital libraries.