Abstract:
The objective of this study is to compare suitable forecasting models for predicting the amount of waste from the packaging process of electronic products. In this study, there is additional interest in the amount of waste from the packaging process between two types of products. Therefore, the data is divided into two groups: 1) Product Type A and 2) Product Type B, with data collected from a case study company for forecasting and comparing the results. The data is monthly, covering the period from January 2019 to December 2024, totaling 72 months, to be used for comparing statistical forecasting methods, including (1) Exponential Smoothing Method, (2) Time Series Regression Method for forecasting with seasonal variation, (3 ) Box-Jenkins Method, and (4 ) Combined Forecasting Method. The accuracy of the forecasting models will be assessed using the Mean Absolute Percentage Error (MAPE). The expected outcome of this study is to use time series analysis techniques for forecasting to identify the most suitable model, which will aid in analyzing the waste rate and finding effective ways to reduce waste in the process