A Two-Phase Approach for Stance Classification in Twitter Using Name Entity Recognition and Term Frequency Feature
Yin Min Tun, Phyu Hninn Myint · 2019
People can trace a course or derivation of subject from the tweet that the speaker is likely against the subject or target. The purpose of the task is to develop the automated systems to determine whether people can infer the twitter. Most of automatic stance classification systems generally need to recognize parts of related information that may not be present in the focus text. And it also needs to decide the tweet whether favor or against the topic depending on the condition if this tweet present the target or focus text. Name Entity Recognition (NER) is the one of the best method to classify the stance for neutral and non-neutral. If the stance is non-neural, the term frequency weighting feature is extracted to classify the tweet whether favors or against the target. The main purpose of this paper is to identify the tweet by using the combination of NLP and text mining based features. System Performance in SemEval 2016 Task A and Task B confirm the success of our approach to classifying the stance in Twitter. Our approach for stance classification is also suitable in other social data and opinion mining problems in social network.