An Ensemble Learning Based Approach for Detecting and Tracking COVID19 Rumors

Rumors regarding epidemic diseases such as COVID 19, medicines and treatments, diagnostic methods public emergencies can have harmful impacts on health political, social other aspects of people’s lives, especially during emergency situations crises. With huge amounts content being posted to media every second these situations, it becomes very difficult detect fake news (rumors) that poses threats the stability sustainability healthcare sector. A rumor is defined a statement for which truthfulness has not been verified. During people found difficulty in obtaining most truthful easily because amount unverified information media. Several applied detecting rumors tracking their sources 19-related information. However, few studies conducted this purpose Arabic language, unique characteristics. Therefore, paper proposes comprehensive approach includes two phases: detection tracking. In phase study carried out, several standalone ensemble machine learning were Arcov-19 dataset. new model was used combined models: The Genetic Algorithm Based Support Vector Machine (that works users’ tweets’ features) stacking method texts). phase, similarity-based techniques obtain top 1% similar tweets target tweet/post, helped find source rumors. experiments showed interesting results terms accuracy, precision, recall F1-Score (the accuracy reached 92.63%), findings ROUGE L similarity techniques.