Parameter Optimization Of Holt – Winters Exponential Smoothing Using Golden Section Method for Predicting Indonesian Car Sales

Indonesia's automobile industry is growing rapidly, these companies have absorbed a lot of Indonesian workers. Of course, this is an advantage for Indonesia because it can reduce unemployment. The sustainability of these factories is very important for Indonesia. Information technology has helped humans, especially in the economic field. In this field, information technology can be used to make sales predictions. The right prediction technique can allow companies to get more revenue and make it easier for companies to plan policies in production. This research proposed integration of Holt-Winters exponential smoothing with golden section parameter optimization for predicting car sales in Indonesia. The golden section method is used to find the optimum parameters as input for the Holt-Winters exponential smoothing method. The parameters obtained will be used to predict car sales in Indonesia. Based on experiments on each car brand, three different parameters were obtained to produce the optimum MAPE value. The Optimization of the Multiplicative Holt-Winters parameter resulted in a MAPE of 17,29128 on the Daihatsu brand data. While the Optimization of Additive Holt-Winters parameter gives MAPE 11.7528 results for Toyota car brand sales data and MAPE 17,29128 results in Honda car brand sales data. The resulting sales data for the Suzuki MAPE brand is 34.25001.