Introduction: With the rapid popularity of the Internet, people are increasingly communicating on the Internet and exchanging opinions on topics of interest or content of online news information. Therefore, the automatic detection of topics in group chat has become increasingly important. This paper is a paper published in PRICAI 2016. It proposes a new topic detection method based on implicit replies. Implicit replies implicitly imply chat information in interactions.
Title: Group chat topic detection based on implicit reply
Abstract: Group chat topic detection has become a very significant research because it is widely used in instant messaging (IM) systems. Previous work has focused mainly on improving text similarity between two related texts by using different weight ratios. However, due to the nature of text messages in group chats, the text's similarity may be zero (or close to zero). In order to solve this problem, this paper proposes a new topic detection method based on implicit replies. Implicit replies implicitly imply chat information in the interaction. The comparative experiment results collected from the QQ group dataset demonstrate that the method performs better than the benchmark method.
Keywords: topic detection; group chat; multi-topic window
The first author introduction:
Xinyu Zhang
Hangzhou Dianzi University Computer College
Via PRICAI
Paper original file download
Lei Feng network (search "Lei Feng network" public concern) by: This article is compiled by Lei Feng network exclusive, without permission to prohibit reprint!
