Forecasting Pertalite Stock Expenditures Using Exponential Smoothing and Linear Regression

Abstract

In the current industrial and business era, effective inventory management is essential for maintaining operational sustainability, particularly in the fuel industry. Pertalite, a popular fuel in Indonesia, with an octane number of 90, offers cleanliness, efficiency, and affordability. However, challenges arise in stock expenditure management due to inaccurate forecasting methods. Data mining, utilizing statistical and machine learning techniques, can identify patterns and trends for better stock forecasting. Recent studies highlight the effectiveness of exponential smoothing and linear regression in fuel demand forecasting. Exponential smoothing, which gives more weight to recent data, improves prediction accuracy, while linear regression analyzes the relationship between fuel stock and various independent variables. This study examines Pertalite fuel sales data from May 2022 to April 2024 from a Pertamina gas station in North Lampung. Results show that linear regression can predict trends, while exponential smoothing, using alpha values between 0.1 and 0.9, captures trends and variations over time. Both methods provide stable forecasts for specific months, demonstrating their utility in understanding Pertalite fuel sales patterns. The study underscores the importance of accurate forecasting in inventory management to meet market demands and maintain operational efficiency.

Keywords

• exponential smoothing, linear regression, pertalite, forecast