Bayesian Network to Identify Comorbidities Correlations under Different Scenarios

The primary purpose of this chapter is to complement our previous work. This chapter scrutinizes, interprets the results, and adds details that we did not mention in our previous study. This chapter presents a probabilistic analysis of the COVID-19 population risk for the different geographical regions of Mexico by comorbidities analysis using the Bayesian Network and a Direct Acyclic Graph.

The research period runs from April 12 to June 29, 2020. (220,667 patients). The method is applied in nature and structured in a descriptive and explanatory statistical analysis type based on its level of involvement in the subject of study. The Mexican dataset used here has quantitative and semi-quantitative characteristics because they result from a questionnaire instrument comprising 34 fields, including the nine comorbidities, hospitalized and non-hospitalized patients, and the virus test.

According to the manipulation of the variables, this chapter describes the input Bayesian Network preprocessing to determine the conditional independent comorbidity occurrence in each geographical region of Mexico.

As a result of this chapter, there is a higher risk of people losing their lives in Region 1 and Region 4 than in Region 2.