A Recommender System For Social Book Search

Abstract

Information retrieval (IR) is a field of computing which deals with storing and retrieving document information. The World Wide Web (WWW) contains a vast amount of information. Storage and retrieval of this information is a huge task. Extensible Markup Language (XML) is used to represent documents so that portions (or elements) may be effectively retrieved. INEX (Initiative for the Evaluation of XML retrieval) is a forum for experimental XML retrieval. It is used to evaluate XML retrieval systems and provides a number of tracks (e.g., Social Book Search, Linked Data, and Tweet Contextualization) and evaluation strategies for the systems designed by competing teams. It also provides a set of XML documents and queries that can be used as a test bed. This thesis focuses on the 2014 INEX Social Book Search (SBS) Suggestion task. The goal of this track is to provide support to users in searching and navigating a large set of books using professional and metadata and user-generated content. In this task, given book requests from LibraryThing discussion forums and a collection of 2.8 million book descriptions from Amazon and LibraryThing, a ranked list of book suggestions is returned to the user. The methodology (based on traditional retrieval and recommendation), the experimental results, and conclusions are described herein.