Tweets Topic Classification and Sentiment Analysis Based on Transformer-Based Language Models

People provide information on their thoughts, perceptions, and activities through a wide range of channels, including social media. The acceptance media results in vast volume valuable data, variety format as well veracity. Analysis such ‘big data’ allows organizations analysts to make better faster decisions. However, this data had be quantified has extracted, which can very challenging because possible ambiguity complexity. To address extraction, many analytic techniques, text mining, machine learning, predictive analytics, diverse natural language processing, have been proposed the literature. Recent advances Natural Language Understanding-based techniques more specifically transformer-based architectures solve sequence-to-sequence modeling tasks while handling long-range dependencies efficiently. In work, we applied sequence short texts’ topic classification sentiment analysis from user-posted tweets. Applicability models is investigated posts Great Barrier Reef tweet dataset obtained findings are encouraging providing insight that for researchers working large datasets number target classes.