This study discussed the modeling enterovirus cases in Taiwan area by using a 10-year/120-month dataset during January 1999 to December 2008 from the surveillance of Centers for Disease Control, Taiwan. This study showed that the most appropriate model was an autoregressive model with order 1 (AR(1)) for Taipei district and North district; an autoregressive model with order 2 (AR(2)) for Central district, South district, and Kaoh-Pin district. These low-order autoregressive models, AR(1) and AR(2) were introduced to represent the nature of the phenomena in the sequence of the enterovirus surveillance data. Furthermore, we analyzed the cross correlation among the neighboring districts. The results showed that there exists a statistically significant positive correlations between the cases of the northern district and the southern district with lagging zero (h=0,1?15days). Public health officials can therefore benefit from the constructed models and mentioned above cross-correlation between the neighboring districts. This information can be used for determining the epidemic control policies or strategies, such as the timing for suspending classes, and epidemic notification through media.
