Hybrid model of ARIMA-linear trend model for tourist arrivals prediction model in Surakarta City, Indonesia

It is important to predict the tourist arrival to help the government in making appropriate decisions. Many models have been proposed to estimate the number of tourist arrivals in the future. An autoregressive integrated moving average (ARIMA) model, linear trend and Holt-Winter triple exponential smoothing are among successful models used in various fields. In the present study, we propose a hybrid model that combines ARIMA and linear trend model as a tourist arrivals prediction model. Experiment results show that the hybrid model produces better prediction performance compared to ARIMA, linear trend and Holt-Winter triple exponential smoothing models.