SARS-Cov bogus news detection: A deep structured learning approach using BERT

The Rapid outbreak of SARS-Cov created total havoc among the people, which led to an epidemic. People around the world started to believe in the information that passes through the internet without any verification, which is the main cause of an infodemic. Some unverified news may have the capability to harm the people in general. To be more aware of the misinformation or unverified news we will be approaching Natural language processing methodologies that use a neural network model, a language model which is trained in advance i.e., BERT. Bidirectional capabilities of this language model BERT help to detect the hidden message from the piece of information. We will be using a benchmarked dataset of covid-19 which is available publicly. Via computing the linguistic conceptual model between both the claimed and factual information obtained out of a carefully selected COVID-19 datasets, this model checks the claim is true or not and it pulled the accuracy of 89%.