Development of a Web-Based Forecasting System Using the Holt-Winters Exponential Smoothing Method to Improve Accuracy in Predicting Cut Flower Harvest Needs

Information is the result of data processing so it becomes important for the recipient and is useful as a basis for decision making. In the flower business, production information is needed to determine the number of cut flowers to match customer demand in the future. This aims to minimize losses that occur due to a shortage or excess of cut flowers. There are still live flower shops that use manual methods and feelings to determine cut flower needs. Therefore, a website-based forecasting system is needed to provide estimated flower cuts for the next period for businessmen. The application was built using the PHP programming language by implementing the Holt-Winters exponential smoothing forecasting method. The Holt-Winters method was used to estimate flower needs, especially cut flowers for bouquet products. Calculation of error score in the method was calculated by MAPE testing. Based on the test results, it is known that the error score in the Holt-Winters forecasting method for the number of cut flowers in a bouquet of Aster, Roses, and Pikok is 13.89%, 16.54%, and 13.23% respectively and they are in Good category. This method has lower error percentage for Aster, Rose, and Pikok flower respectively by 3.95%, 1.86%, and 14.01% than by using florist former method.