Overview of CryptOQA: Opinion Extraction and Question Answering from CryptoCurrency-Related Tweets and Reddit Posts
Gourav Sen, Sougata Sarkar, Subham Kumar, Koustav Rudra, Kripabandhu Ghosh · 2024
Cryptocurrency is an exponentially growing domain, and Twitter and Reddit are important outlets for users to hold opinions or ask questions or discuss about certain topics.The CryptoQA track at FIRE 2024, seeks to develop systems which will automatically assess cryptocurrency posts on social media.In this track, there are basically two tasks: (a) the classification of posts related to cryptocurrencies into the eight classes namely, Noise, Objective, Positive, Negative, Neutral, Question, Advertisement, Miscellaneous divided across three levels and (b) the participants had to detect whether the answer is relevant to the question given question-answer pair.Overall, three teams from various institution have submitted to this track.These experiments were evaluated using some quantitative measures, particularly the F1-Score, and a qualitative evaluation based on some features developed, supervised learning models and their learning parameters.These models were able to successfully tackle the issues of sentiment dispersion, noise, and several versions of posts which helped a great deal in grasping the lingering noise from the crowd in the context of cryptocurrency.