The goal of this study was to propose modeling strategies applied to the analysis of citrus leprosis incidence, through the use of a spatial temporal autologistic model. We evaluated the adequacy of autologistic model to consider data collected at different times; to detect spatial-temporal patterns through different neighboring structures; to consider the effect of covariates from previous times; and assessing the effect of the presence of the disease vector in the probability of new infections occurrence. The spatial temporal autologistic model adopted has extended the usual logistic model, in which the neighboring structures is described by means of covariates built from the status of plants nearby, at the same or at previous times. Data regarding the presence of the leprosis on plants were collected at field points referenced in space, over a period of approximately two years. Models detect the presence of spatial patterns on new infections for the studied neighboring structures, at the same or previous time. Additionally, probability estimates of a plant become infected can be obtained from the fitted models, given the occurrence of the disease and vector.
spatial statistics; neighbor structures; pseudo-likelihood
