Abstract:
This research aimed to find the most suitable technique for forecasting the usage of three items namely S41123389000 S41123429000 and S41231283000. The data used in this study was 24 months based on past usage. The comparative approach was applied in order to choose the most appropriate forecasting method that is applicable to the data from each item. Due to actual usage not have trend and seasonal, the three methods chosen for comparison are 1) Moving Average Method 2) Single Exponential Smoothing Method and 3) Linear Trend Line Method. Mean Absolute Percentage Error (MAPE) was used as indicator to choose the best of all three methods. As the research reveals, the data analysis of S41123389000 shows the lowest MAPE when Moving Average Method applied, while the results of S41123429000 and S41231283000 are at best when using Linear Trend Line Method. Furthermore, the results of forecasting of the three items by the above chosen methods are more precise in value than the ones using currently by the company by 48.21% 46.55% and 18.47 percent respectively