การประยกตใชตวแบบการถดถอยในตวอนดบทหนงทมคาเปนจำนวนเตมในการพยากรณจำนวนอบตเหตของรถบรรทก จงหวดสรนทร

This study aims to forecast the monthly number of truck-related accidents in Surin Province. The accident data are non-negative integers, as time series count data, which differ from conventional time series that are typically real numbers. In the presence of a large number of zeros, models suitable for zero-inflated data may be considered. Therefore, this study employed the First-order Integer-valued Autoregressive (INAR1) model and then compared it with the Zero-inflated Poisson INAR1 (ZIPINAR1) model and the Zero-inflated Negative Binomial INAR1 (ZINBINAR1) model. Model selection was based on the Akaike Information Criterion (AIC). The dataset comprised 72 monthly observations of truck accidents in Surin Province from January 2019 to December 2024. The results indicated that the ZIPINAR1 model provided the best fit, yielding the lowest AIC value of 183.6693 compared to the other models.