ComStreamClust: a Communicative Multi-Agent Approach to Text Clustering in Streaming Data

Topic detection is the task of determining and tracking hot topics in social media. Twitter arguably most popular platform for people to share their ideas with others about different issues. One such prevalent issue COVID-19 pandemic. Detecting on these kinds issues would help governments healthcare companies deal this phenomenon. In paper, we propose a novel, multi-agent, communicative clustering approach, so-called ComStreamClust sub-topics inside broader topic, e.g., FA CUP. The proposed approach parallelizable, can simultaneously handle several data-point. LaBSE sentence embedding used measure semantic similarity between two tweets. has been evaluated by metrics as keyword precision, recall, topic recall. Based recall number keywords, obtains superior results when compared existing methods.