INTRODUÇÃO: A malária é uma doença endêmica na Amazônia Legal Brasileira, apresentando riscos diferentes para cada região. O Município de Cantá, no Estado de Roraima, apresentou para todo o período estudado, um dos maiores índices parasitários anuais do Brasil, com valor sempre maior que 50. O presente estudo visa à utilização de uma rede neural artificial para previsão da incidência da malária nesse município, a fim de auxiliar os coordenadores de saúde no planejamento e gestão dos recursos. MÉTODOS: Os dados foram coletados no site do Ministério da Saúde, SIVEP - Malária entre 2003 e 2009. Estruturou-se uma rede neural artificial com três neurônios na camada de entrada, duas camadas intermediárias e uma camada de saída com um neurônio. A função de ativação foi à sigmoide. No treinamento, utilizou-se o método backpropagation, com taxa de aprendizado de 0,05 e momentum 0,01. O critério de parada foi atingir 20.000 ciclos ou uma meta de 0,001. Os dados de 2003 a 2008 foram utilizados para treinamento e validação. Comparam-se os resultados com os de um modelo de regressão logística. RESULTADOS: Os resultados para todos os períodos previstos mostraram-se que as redes neurais artificiais obtiveram um menor erro quadrático médio e erro absoluto quando comparado com o modelo de regressão para o ano de 2009. CONCLUSÕES: A rede neural artificial se mostrou adequada para um sistema de previsão de malária no município estudado, determinando com pequenos erros absolutos os valores preditivos, quando comparados ao modelo de regressão logística e aos valores reais.

INTRODUCTION: Malaria is endemic in the Brazilian Amazon region, with different risks for each region. The City of Cantá, State of Roraima, presented one of the largest annual parasite indices in Brazil for the entire study period, with a value always greater than 50. The present study aimed to use an artificial neural network to predict the incidence of malaria in this city in order to assist health coordinators in planning and managing resources. METHODS: Data were collected on the website of the Ministry of Health, SIVEP - Malaria between 2003 and 2009. An artificial neural network was structured with three neurons in the input layer, two intermediate layers and an output layer with one neuron. A sigmoid activation function was used. In training, the backpropagation method was used, with a learning rate of 0.05 and momentum of 0.01. The stopping criterion was to reach 20,000 cycles or a target of 0.001. The data from 2003 to 2008 were used for training and validation. The results were compared with those from a logistic regression model. RESULTS: The results for all periods provided showed that the artificial neural network had a smaller mean square error and absolute error compared with the regression model for the year 2009. CONCLUSIONS: The artificial neural network proved to be adequate for a malaria forecasting system in the city studied, determining smaller predictive values with absolute errors compared to the logistic regression model and the actual values.

Roraima is the northernmost state of Brazil, bordering both Venezuela and Guyana. Appropriate climate and vector conditions for dengue transmission together with its proximity to countries where all four dengue serotypes circulate make this state, particularly the capital Boa Vista, strategically important for dengue surveillance in Brazil. Nonetheless, few studies have addressed the population dynamics of Aedes aegypti in Boa Vista. In this study, we report temporal and spatial variations in Ae. aegypti population density using ovitraps in two highly populated neighbourhoods; Centro and Tancredo Neves. In three out of six surveys, Ae. aegypti was present in more than 80% of the sites visited. High presence levels of this mosquito suggest ubiquitous human exposure to the vector, at least during part of the year. The highest infestation rates occurred during the peak of the rainy seasons, but a large presence was also observed during the early dry season (although with more variation among years). Spatial distribution of positive houses changed from a sparse and local pattern to a very dense pattern during the dry-wet season transition. These results suggest that the risk of dengue transmission and the potential for the new serotype invasions are high for most of the year.