Comparison of Multiple Linear Regression and Holt-Winter Exponential Smoothing in the Gold Jewelry Pricing Prediction

Abstract

One important factor in someone's decision to purchase gold jewelry is the price factor. However, the price of gold jewelry varies from store to store. Therefore, a gold jewelry price forecasting system is needed to assist customers in determining the right time to buy gold jewelry. This study aims to develop a gold jewelry prediction system by comparing two methods: multiple linear regression and Holt-Winter exponential smoothing. These two methods were chosen because the researcher wants to compare which method is more accurate in predicting the price of gold jewelry. The second reason is that both multiple linear regression and Holt-Winter exponential smoothing methods share similarities in using three variables. The three variables in multiple linear regression include period (X1), profit (X2), and price (Y), while the variables in the Holt-Winter exponential smoothing method include level, trend, and season.