出版社：Grupo de Pesquisa Metodologias em Ensino e Aprendizagem em Ciências
摘要：Dengue is one of the serious public health problems worldwide. The Northeast of Brazil (NEB) has an ideal climate and urban environment for the proliferation of the Aedes mosquito (aegypti and albopictus), the vector of the disease. The State of Alagoas, especially its capital, has frequent epidemics of the disease. Therefore, the objective of this study is to evaluate the application of Artificial Neural Networks (ANN) in reported cases of dengue (CND) in the administrative regions (AR) of Maceió. The RAs are divided into: AR1, AR2, AR3, AR4, AR5, AR6, AR7 and AR8. The CND were submitted to ANN nonlinear autoregressive (NAR) – (ANN-NAR). The study period was 2011 to 2020. The results obtained from the CND stood out in specific years (2012, 2013, 2017, 2018 and 2020), on the other hand, there were overestimations of the forecasts via ANN. In some ARs there was underreporting and, therefore, it interfered with the forecasts results. The ANN-NAR was validated, as most of the predictions showed a positive correlation and responded to the observed data, except for the ARs with underreporting. The use of ANN is suitable for warning and disease prediction, where such an instrument can be used in preventive actions to control the disease.
关键词：Dengue;Northeast of Brazil;Artificial intelligence;Urban climate.
