Modelling tourists arrival using time varying parameter

International Conference on Mathematics: Education, Theory and Application

数学;教育

Suciptawati, P.^1 ; Sukarsa, K.G.^1 ; Kencana, Eka N.^1

Department of Mathematics, Udayana University, Kampus Bukit Jimbaran-Badung, 80361, Indonesia^1

关键词: Dependent variables;    Economic development;    Kalman filter algorithms;    Mean absolute percentage error;    Out-of-sample forecast;    Poverty reduction;    Predictor variables;    Time varying parameter;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/855/1/012047/pdf
DOI  :  10.1088/1742-6596/855/1/012047

学科分类：发展心理学和教育心理学

来源: IOP

【 摘 要 】

The importance of tourism and its related sectors to support economic development and poverty reduction in many countries increase researchers' attentions to study and model tourists' arrival. This work is aimed to demonstrate time varying parameter (TVP) technique to model the arrival of Korean's tourists to Bali. The number of Korean tourists whom visiting Bali for period January 2010 to December 2015 were used to model the number of Korean's tourists to Bali (KOR) as dependent variable. The predictors are the exchange rate of Won to IDR (WON), the inflation rate in Korea (INFKR), and the inflation rate in Indonesia (INFID). Observing tourists visit to Bali tend to fluctuate by their nationality, then the model was built by applying TVP and its parameters were approximated using Kalman Filter algorithm. The results showed all of predictor variables (WON, INFKR, INFID) significantly affect KOR. For in-sample and out-of-sample forecast with ARIMA's forecasted values for the predictors, TVP model gave mean absolute percentage error (MAPE) as much as 11.24 percent and 12.86 percent, respectively.