Forecasting Spare Parts Demand of Military Aircraft: Comparisons of Data Mining Techniques and Managerial Features from the Case of South Korea

摘要: In a weapon system, the accurate forecasting of spare parts demand can help avoid excess inventory, leading to efficient use budget. It also develop combat readiness system by improving utilization. Moreover, as performance-based logistics (PBL) projects have recently emerged, has become an important issue for PBL contractors well. However, parts, time series methods, typically used in military sector, low prediction accuracies and are mostly based on judgment practitioners. Meanwhile, most previous studies sector not considered managerial characteristics (e.g., reparability irregularity maintenance). No work any such features, which indicate reliability mean between failures (MTBF)), although they affect demand. Therefore, more aircraft, we designed examined systematic approach that uses data mining techniques. To fill up research gaps related works, our included new features represent parts. Consequently, given case South Korea full feature set, found random forest gave better results than other techniques conventional methods. Using best technique Random Forest, identified contribution each set accuracy, operation environment valuable sets significant way, so should be collected, managed carefully, aircraft.