Stance Detection using Two Popular Benchmarks: A Survey

Abstract

The task of extracting knowledge from text is a significant process of many intelligent applications. Stance detection is one of the strongly coming learning-based objectives. Especially, on social media (SM) available huge text. Since SN content is increasing massively every second. The tasks of stance detection (StD), like any learning-based process, need to train and test on a robust pre-made dataset benchmark. The English language has the most available benchmarks sources. Occasionally, an StD researcher needs to gain deep knowledge about these researches involving a particular domain before establishing research scope and objectives. This study aimed to help StD researchers in this emerged field; i.e. StD. A focused review on a specific domain is presented. Two StD benchmarks on English are determined as the domain of StD surveyed papers. StD Challenges and problems are introduced beside the StD models investigation. This survey’s first goal is a point to the stance detection concept and the two popular benchmarks. The second is to discuss StD involved models and managed problems.