Machine Learning as a Method for Reclassification of Deaths from Unspecified Cerebrovascular Disease in Brazil, 2000–2019

This dataset contains individual-level mortality records for deaths attributed to unspecified stroke (US) in Brazil, an upper-middle-income country, between 2000 and 2019. The data were extracted from the Mortality Information System (SIM) according to ICD-10 coding criteria. Mentions of cerebrovascular disease were identified in lines A–D and part 2 of death certificates.

This dataset supports research on cause-of-death classification, the application of artificial intelligence in epidemiology, and the study of cerebrovascular disease burden in Brazil.