A detection method of FAQ matching inquiry e-mails by automatic generation of characteristic word groups on past inquiry e-mails

This paper discusses how to detect the inquiry e-mails corresponding to pre-defined FAQs(Frequently Asked Questions). Web-based interaction such as orders and registration form on a web page is usually provided with its FAQ page for helping a user, however, most users submit their inquiry e-mails without checking such pages. Help desk operators must spend lots of time on replying to process lots of e-mails even if some contents match to FAQs, we propose the automatic detection method of such e-mails based on SVM (Support Vector Machine) and specific Jaccard coefficient with positive and negative instances. Experimental results show that the proposed method can improve the recall rate to 50% and the precision rate to 91%.