Early Warning Panel

Resumo

Artificial intelligence methods are normally used to make predictions, classify objectives, group and optimize human work and minimize uncertainty for decision making, but despite the advantages they offer, they are not entirely friendly or accessible to the majority of people. That is why we use an artificial intelligence model, the LSTM model, it is used to predict time series and can be extended to any endemic disease for which there is enough data for deep learning of an artificial intelligence model, which which will allow us to draw up plans to combat and confront this disease in our country.
Paraguay is a country that epidemiologically is considered endemic for some arboviruses, we can mention: Dengue with cases since 2009, Chikungunya since 2013 and Zika since 2015.
The epidemiological indicators provide us with information on the epidemiological situation of Dengue in Paragua, the indicators that we collect through a bibliographic review and that found in our Early Warning Panel were identified through a bibliographic review, the search engine used was Google Scholar, the inclusion criteria were: articles that mention epidemiological indicators and the way in which they are calculated. Finally, the epidemiological indicators used were: Incidence, bitting rate, probability of infection from mosquito to human, probability of infection from human to mosquito.
For the implementation of the prediction in our Early Warning Panel, the LSTM model was defined, pre-training. It is one of the most advanced and successful deep learning architectures for time series prediction.
For the evaluation of our Panel, a survey was developed for potential users in order to evaluate the usability, relevance and performance of the system. Experts in epidemiology and control of infectious diseases, with experience in managing dengue cases, were selected. A representative sample was sought that included epidemiologists, doctors specializing in tropical diseases, data analysts, and public health officials. With the potenticial users was evaluated: Usability, Relevance and Performance.
The results of this evaluation revealed good acceptance by the experts, who highlighted the clarity and usefulness of the information presented. These findings confirm the effectiveness and relevance of the epidemiological panel as a useful tool for the surveillance and control of dengue.
The panel is available at: https://alertatemprana.vercel.app