Learning from message pairs for automatic email answering

Abstract

We consider the problem of learning a mapping from question to answer messages. The training data for this problem consist of pairs of messages that have been received and sent in the past. We formulate the problem setting, discuss appropriate performance metrics, develop a solution and describe two baseline methods for comparison. We present a case study based on emails received and answered by the service center of a large online store. © Springer-Verlag Berlin Heidelberg 2004.