A Semi-Supervised Approach to Detecting Stance in Tweets

TL;DR

This paper introduces a semi-supervised method for stance detection in tweets, leveraging hashtag-based data collection and dependency features to classify pro or con opinions on various social issues.

Contribution

It proposes a novel semi-supervised approach that uses high-precision hashtags to create large training datasets for stance detection without manual labeling.

Findings

Effective stance classification on Twitter data.

High precision hashtags improve training data quality.

Dependency features with sentiment lexicons enhance performance.

Abstract

Stance classification aims to identify, for a particular issue under discussion, whether the speaker or author of a conversational turn has Pro (Favor) or Con (Against) stance on the issue. Detecting stance in tweets is a new task proposed for SemEval-2016 Task6, involving predicting stance for a dataset of tweets on the topics of abortion, atheism, climate change, feminism and Hillary Clinton. Given the small size of the dataset, our team created our own topic-specific training corpus by developing a set of high precision hashtags for each topic that were used to query the twitter API, with the aim of developing a large training corpus without additional human labeling of tweets for stance. The hashtags selected for each topic were predicted to be stance-bearing on their own. Experimental results demonstrate good performance for our features for opinion-target pairs based on generalizing dependency features using sentiment lexicons.