Comparison of double exponential smoothing and triple exponential smoothing methods in predicting income of local water company

Abstract

Local Water Company is a government-owned business entity that has a business scope in the management of drinking water and water facilities to improve the welfare of the community, which includes social, health, and public service aspects and has a very important role for the community in terms of supporting the smooth development of the region so that the company's success must always be sought. This effort certainly needs to be supported also in terms of setting revenue targets at the company. Determination of revenue targets in local water companies is currently using manual calculation methods so that the accuracy and effectiveness of setting revenue targets is less accurate. Forecasting models with mathematical methods are needed to predict future revenue targets so that monitoring of the success of regional development and consideration in decision making can be monitored. Forecasting this revenue target is based on actual data within the previous 5 (five) years, namely from January 2014 until December 2018, by comparing the two forecasting methods, namely the Double Exponential Smoothing (DES) and Triple Exponential Smoothing (TES) methods. The forecasting accuracy method is used the Mean Absolute Percentage Error (MAPE) method to measure the accuracy of the forecasting results from the two forecasting methods used. Forecasting test results are performed using alpha constant values of 0.1, 0.3, 0.5, 0.7, and 0.9 as trial data. Based on the trials conducted, the forecasting results presented show that forecasting results with the Double Exponential Smoothing method provide a more optimal forecasting result at alpha 0.7 with a MAPE value of 9.54%, so that the use of the Double Exponential Smoothing method is recommended in forecasting revenue targets in the Local Water Company because it has an error value under 10%.