Linking, mining and visualising large textual datasets : the PQ use case

Abstract

This paper addresses the general problem of access to public data. We propose a systematic approach which, in contrast to systems currently available, will provide effective access to large quantities of public information in a form that makes sense to the vast majority of ordinary citizens. The key idea is to use current semantic and human language technologies for the fully automatic extraction of high level information from large textual datasets. This can be used to classify, organise, and relate the contents of documents in a highly flexible way that can be personalised to suit the needs of the individual user. We describe an experimental system that is being developed to handle a dataset of 120,000 Parliamentary Questions.