Abstract

ALDAZ CARDENAS, J.W; GARCIA DIAZ, J.R  and  QUINONES RAMOS, R. Canine parvovirus in the Bolivar province, Ecuador. Utility of Box-Jenkins models for its analysis and prediction. Rev Salud Anim. [online]. 2012, vol.34, n.3, pp. 165-172. ISSN 0253-570X.

The objective of the present work was to evaluate the prevalence and mortality of canine parvovirus and the feasibility of the model ARIMA (Box-Jenkins) for its analysis and prediction. Retrospective data was obtained from the veterinary hospital «Caninos y Felinos» of the town of Guaranda; The monthly prevalence and mortality of the disease was evaluated and series of time were defined with the values obtained to which a Box-Jenkins model was fitted determining its parameters to identify mixed models with ARMA structure. The models fitted to the series were used to prognosticate the values of the year 2011, comparing with the values observed through the Kolmogorov-Smirnov test.The prevalence and mortality showed a semiannual seasonal behavior; the model fitted for prevalencewas ARIMA (2,0,2)x(1,0,1)6 with constantand for mortality ARIMA(1,0,0)x(2,0,2)6 with constant. In each case, the same behavior pattern was prognosticated for the next 12 months with a high correspondence between the real data and those prognosticated with the models (P>0,010). It is concluded that the prevalence and the mortality has a six-monthly seasonal pattern and can be modulated and prognosticated with high precision through Box-Jenkins models making easy its application in these and other indicators of animal health always that the observations can be expressed as a series of time.

Keywords : Canine parvovirus; series of time; seasonal behavior; prognosis.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )