A Personalisable Hierarchical Intranet Document Categoriser (MSc Thesis)


A novel system is proposed for the indexing, searching and browsing of an intranet document repository for use as part of a corporate extranet. The system allows users to browse a hierarchically organised collection of documents. The hierarchy is automatically maintained by the system after a minimum of training. Users additionally have the option to personalise the hierarchy in order to organise the documents in any way they see fit. This document presents an overview of the design of the system and the reasons for the various design choices.

The system has been implemented and initial tests on the system have been conducted which show that the system would comfortably be able to handle a repository sized between 3,000 and 30,000 documents.

The system was developed with support from EDF.

Full Report

The project report is available here.


Presentation slides are available here.

Project Information

Launch Date: October 2005
Project Language: C#
Start Date: March 2005