ANALYSIS OF MOVING AVERAGE AND HOLT-WINTERS OPTIMIZATION BY USING GOLDEN SECTION FOR RITASE FORECASTING

摘要

Moving average and exponential smoothing holt winters is a method for forecasting calculations of some existing statistical and computational method. The selection of two methods in this research is based on ritase data from January 2013 December 2015 to view the average, trend and seasonal pattern that will occur in one period of future forecasting research. However, the weakness from Exponential Smoothing Holt Winters is the determination for initial values (α, β, & γ) are inputted with trial value from 0 to 1 which may not yield maximum results. Golden section method is added in this research to assist the optimum determination for initial values (α, β, & γ) from Exponential Smoothing holt Winters to produce the accurate results. This research aims to know the forecasting model for the amount of income ritase at Department of transportation Yogyakarta UPT Giwangan terminal management with Exponential Smoothing holt Winter’s and moving average. Furthermore, to know the comparison of forecasting results with both methods. To obtain the right method, the measuring instrument is needed to detect the accuracy of the prediction value, while the one used in this research is the mean absolute percentage error (MAPE), mean square deviation (MSD) and Mean Absolute Deviation (MAD).The determination of forecast value and selection of MAPE, MSD, and smallest MAD is using the two methods above. The results of data analysis shows that Exponential Smoothing holt Winters is considered as the right method for the amount of ritase income at Department of Transportation, city of Yogyakarta - UPT Management of Giwangan Terminal because it produces the smallest value of MAPE = 4%, MSD = 446841 and MAD = 496.