BETANCOURT BETANCOURT, José A; SANTANA BRITO, Humberto; ORTIZ HERNANDEZ, Eloy    RODRIGUEZ SOCARRAS, Nuria. Characterization and analysis of acute respiratory diseases time series from Camagüey. []. , 13, 1, pp. 0-0. ISSN 1025-0255.

Background: In Cuba an excellent system of surveillance exists, but has to face the challenge of constant demographic changes and populations' movements inside and outside of the country, among other. An efficient system of alert and answer against the presentation of diseases exists, but these programs may take into account the budget for the public health activities and the complexity of the propagation; the use of mathematical models could support this effort. Objective: To characterize a series of time of acute respiratory diseases in Camagüey province and to evaluate it with an appropriate method to its behavior. Methods: Rates of series of time weekly of acute respiratory diseases were used, reported between the years 2000 to 2007 of Camagüey province, Cuba. To characterize these series Matlab V-7,4-2006 program was used. The stationarity was evaluated with the simple autocorrelation method, the linearity behavior was carried out with the test designed by Brock, Dechert and Scheinkma, with the method based on the spectral geometry that was determined if the series was stochastic or determinist. Errors were compared in the predictions with ARMA and GARCH models, errors and prediction curves were analyzed. Results: The non stationary behavior of the series is demonstrated, based on the simple autocorrelation function. The test indicates that the studied series was not lineal. The spectral Geometry test concluded that the series comes from a stochastic system. In this stationary, not lineal and stochastic series was more adequate GARCH model than ARMA, given the smallest error and more adjustment of the prediction curve. These results demonstrated the necessity to classify the series of time before being analyzed, to propose the most appropriate model. Conclusions: The studied series turned out to be not lineal, no stationary and stochastic and the prediction was superior in this case, with the model GARCH compared with ARMA.

: Epidemic prediction; epidemic time series; characterization time series; model GARCH; epidemiology complex systems.

        · |     · |     · ( pdf )