Stock Forecasting Information System using the Holt-Winters Method

Abstract

In sales, prediction is an important element carried out by traders to help their sales efforts. Predictions involve assessing future values based on historical data. Problems that can arise from the combination of stock and transactions. The aim of this research is to create a sales prediction system, monitor sales and input transactions. This research uses the Holt-Winters method to predict sales. The main goal is to implement the findings of this research into a web administration site. In this research, transaction data is used as a basis for predictions using the Holt-Winters method. The Holt-Winters method uses a series of actual values for predictions. Parameters like Alpha, Gamma, and Beta are defined for weighting, and initial values are calculated for Trend, Level, and Seasonality attributes to achieve predicted results. Evaluation of the Holt-Winters method in this research involves searching for the Mean Absolute Percentage Error (MAPE) to determine the percentage of prediction error. The application of this research involves creating a website administration to assist sellers in combining transactions and available products. The average results of the evaluation of the administration of the website created resulted in a satisfaction percentage of 82.61%. This shows that the system can help in unifying transactions and products.