Prospects for natural language engineering and knowledge discovery

We outline an approach which aims to link data mining techniques within an architecture to assist in understanding natural language texts. It is obvious that understanding language is a kind of knowledge problem, and it is generally acknowledged that knowledge acquisition is costly and time consuming. We suggest that rule induction, and related approaches, can help make this particular problem more tractable, paving the way for various useful and usable products. It is taken as axiomatic that the information, and especially textual information, which is available to individuals and organisations will continue to grow. Unfortunately, the capacity of people to deal with information unaided is going to remain static. Therefore there is a need for tools which can summarise, categorise and contextualise information.