Risk prediction of Covid-19 Vaccine Adverse event by using Bayesian network

the most effective way to against COVID-19 spreading is vaccines. The vaccines that is widespread in any countries are Pfizer-BioNTech, Moderna, Janssen (Johnson & Johnson), etc. However, technology behind to develop the vaccines are also different such as mRNA, Viral vector, and In activated vaccine. Too early launching the vaccines would affect side effect and adverse events depends on patients. This paper exhibit Bayesian Network to construct the model for predicting the adverse event. The aim of this research is to calculate probabilities of adverse event in based case (considering only the dataset) and predict the probabilities in observed nodes from particular scenarios. A previous study reveals the medical idiom in the Bayesian network as a useful instrument to construct the model. In fact, cause-consequence idiom is the most utilized in the graphical network. In terms of methodology, this paper would deploy pybbn framework to elicit the model and using NetworkX to visualise the model. Furthermore, the dataset is extracted from VAERs of CDC and data consist of the adverse event report in United State (Pfizer, Moderna, and Johnson & Johnson vaccines). The analysis from this model reveals that thepossibility of patients who are only got first dose from Pfizer and Moderna is likely to die if patients have severe comorbidity or complications. Whilst probability of death case in Johnson & Johnson is lower than a half. All vaccine brands would be positively affect to reduce death event due to infect COVID-19. In further research, the author would expand the Bayesian network model, including other symptom (complications), comorbidity or chronic disease, and allergies