Lifelong Knowledge Acquisition with Topic Monitoring

Abstract

Industry workers and their R&D colleagues rely on knowledge stored in patents, reports, workplans, guidelines and norms. With time, the subjects considered but also the terminology used in such documents change. This impeds the categorisation of old documents and reduces the effectiveness of keyword-based search. We present a document stream mining solution that organises document archives into topics and monitors the changes of the topics and their characteristic keywords over time.