Cognitive-Linguistics-based Request Answer System

erschienen September 2009 in: Madrid, Spain AMR 2009 - 7th International Workshop on Adaptive Multimedia Retrieval

Closed domains pose very specific problems to applications of all kinds. While e.g. search applications in open domains can access a huge pool of diverse information (e.g. in the internet), this is not possible for closed domains. An example application for this problem are customer support systems. These systems normally administrated by humans have to cope with different types of requests, e.g. a user asking not only a single question, but also giving a description of an experienced situation. Analyzing those complex requests is very diffcult in general. Therefore it is nearly impossible to handle for common search, question-answer or customer support systems. In this paper we propose a system which should be capable of (semi-) automatically analyzing and reacting to different types of requests in closed domains based on a cognitive linguistics approach.