Text document clustering using self organizing map: Theses and dissertations of universitas Indonesia

Abstract

Accessibility is a critical aspect to be considered by college library in order to facilitate users in searching library collections. The Library of Universitas Indonesia, as one of Asia's largest library with more than 1,500,000 book collections, should also concern about accessibility to balance its numerous collections. UI-ana collections or works produced by and associated with Universitas Indonesia; in particular theses (undergraduate and graduate theses) and dissertations are one of the largest numbers of collections in Universitas Indonesia's Library. However, the current collection's management system was still based on the submission of the collection in Universitas Indonesia's Library. Since these collections are arranged with no exact criterion, it is harder for users to find theses and dissertations with the same topic. Therefore, management of these collections based on certain criterion is extremely needed to facilitate users in searching these collections. This research aims to determine the categories that can represent theses and dissertations through abstract text mining of each collection in 2005–2015 with a clustering algorithm, namely Self-organizing Map. This study found 139 categories which will be used to classify theses and dissertations of Universitas Indonesia.