Utilização do algoritmo Apriori para traçar o perfil sociodemográfico do homem brasileiro com câncer de próstata

Abstract

Among the diseases that affect the male population, prostate cancer has increased the mortality rate among them, where it is the sixth malignant neoplasm in the world and in Brazil the first. It becomes a matter of public health that concerns, but if discovered early, the cure rate is high. Despite the initiatives to help the male population against prostate cancer, there is still a lack of guidance regarding diagnosis and treatment. But the initiatives would be better targeted if they had the profiles of patients assisted by them, but it is still a field of research with gaps. In addition, data that can help are stored in large databases with a lot of information, mainly due to the computerization process of the health sector, which makes manual analysis of this data difficult. This work aims to determine the profile of men who are prone to prostate cancer through the Apriori algorithm for the formation of association rules in the Brazilian context. With that, we applied the Apriori algorithm to the INCA database in order to have the association rules. In the end, we realized that the factors of smoking, alcoholism, race and marital status are the factors that stood out the most for appearing in the rules with the highest levels of confidence. However, we infer that the brown race has a higher incidence of prostate cancer in Brazil. Despite the incompleteness of the optional data in the INCA database, it is important to highlight that the analysis was carried out at the national level and can be used to guide campaigns in the context of men's health.