Hidden Patterns in COVID-19 in Mexico: Data Mining with Association Rules

Description

Mexico suffered one of the worst sanitary crises in its recent history —over 7-6 million confirmed cases, hundreds of thousands of officially reported deaths, and an excess mortality rate that surpassed over 830,000 lives. The real story is not only in the official numbers. It is hidden among the millions of clinical records, in combinations even traditional statistics cannot see, patterns reveal how the pandemic took a heavy toll on those already living with diabetes, hypertension, obesity and chronic pulmonary diseases. Hidden patterns in COVID-19 in Mexico is one of the first books that use advanced data mining –specifically association rules– to uncover the invisible connections in the 2023 SISVER national database (over one million records). With algorithms that generate rules with up to 99% confidence and metrics that measure real impact (lift, conviction, deployability), these combinations reveal realities that no descriptive analyst had captured with such precision: laboratory deterioration + ICU + COPD → intubation almost inevitable; diabetes + hypertension obesity exacerbated on the northern border; frail elderly people at increased risk during the winter.