Automatic Question Answering System for Consumer Products

Abstract

In this paper, we present a novel automatic question answering system being the first research to deal with various types of user questions in consumer electronics domain. We gathered more than 185,000 actual data pairs and 50,000 data triples from Samsungs official web site and a crowd question answering web site in order to provide proper answers to user queries. With this system, users can resolve their problems in using consumer electronic gadgets at any time with a guaranteed response time and answer quality by simply asking questions to the system in a natural language form. The system architecture and test results are addressed in the paper.