Stance Mining for Online Debate Posts Using Part-of-Speech (POS) Tags Frequency

Abstract

Online social media have immense data day after day because online users to connect each other, build their community, share their attitudes and publish their opinions. The users have become an important source of content. One of the most social media platforms is online debate forums which allow the users to express their attitude and their feelings towards the different issues. The debate post is informal language and non-standard expressions are very used, and many spelling errors are found due to absence of correctness verification. Our goal is to use the results of applying stance mining in public area the attempt at automatically collecting user's attitudes from the political debate forum where opinions towards public issues are found for government or political organizations to make decisions. This paper presents the linguistic features combining the part-of-speech (POS) tagging features with tf-idf weights different from ordinary features in stance classification and gets a good accuracy.