Best Time Series Model For Elotex Demand Forecasting

Inventory management is carried out to ensure the accuracy of raw material stock in the warehouse. In a chemical raw material distribution company, stockpiling or shortages of raw materials often occur due to fluctuating customer demand. The company is at risk of indirect losses if the product is not sold immediately or if it becomes unavailable. When products are not sold promptly, there is a potential loss due to the limited shelf life of the goods. On the other hand, when products are not available, the company risks losing its customers. The objective of this study is to design a time series model to predict the quantity of chemical raw materials by comparing the accuracy of the Moving Average, ARIMA, and ARMA models. The comparison results will be based on historical demand data for one of the company's products. The product selected in this study is the chemical raw material Elotex, which has the highest demand. The sample data used spans from 2015 to 2023 in daily units. The selection of the best method in this study is determined by considering the model with the lowest RMSE (Root Mean Square Error) value. The research results show that the RMSE value for the Moving Average (MA) model is 3052.7560, the ARIMA model is 4247.9554, and the ARMA model is 4241.8059. Thus, the Moving Average (MA) model, having the lowest RMSE value, is the most accurate model for forecasting the purchase of Elotex chemical raw materials.