Calendar variation model for ticket sales forecasting at Kayangan Port, East Lombok

Abstract

The forecasting of ferry ticket sales aims to prepare in case of a surge in passengers at the Kayangan port, East Lombok in the following years, especially before Eid al-Fitr. The purpose of this study was to determine the calendar variation model that represents the sales volume pattern of class IVA vehicles so that the model can be used to predict the number of class IVA vehicles in ticket sales and to get the forecasting results for January 2019 – December 2019. Calendar variation is a recurring pattern with varying lengths of periods due to the influence of different calendars each year. Eid al-Fitr is one of the examples of the calendar variations that occur in Indonesia because Eid al-Fitr always has a shift in the Gregorian calendar. This shift is caused by the difference between one year in the Hijri calendar dan one year in the Gregorian calendar. These periodic changes give a calendar variation. Affected data by Eid al-Fitr such as the sales volume of the ticket at the ferry port will also have a calendar variation effect. Thus, to analyze the sales volume of tickets used ARIMA method with calendar variation effect or well known as calendar variation method. Based on the analysis, it is known that the significant dummy variable is D1 which is the variable of the month of Eid Al-Fitr. Then conducted ARIMA modeling of the dummy regression residuals. The best model then obtained by analysis, which is the calendar variation model that containing ARIMA (0,1,1)(1,0,0)12, all the parameters are significant and the residual assumption is fulfilled, which are normality test and white noise test. The Mean Absolute Percentage Error (MAPE) of the calendar variation model that containing ARIMA (0,1,1)(1,0,0)12 is 12,6% which means that the forecasting results can be trusted by 87,4%.