Machine learning-based hybrid preprocessing techniques for UAV spare parts demand forecasting

Recently, there is been a surge in interest in unmanned aircraft as strategic tools for Defense Readiness Condition (DEFCON). In accordance with this worldwide trend, the Korean military has developed unmanned aerial vehicles (UAV) in an effort to improve DEFCON. To ensure the proper operation of these vehicles, it is important to accurately forecast the demand for spare parts for equipment maintenance and procurement. In order to forecast the demand for spare parts, the Korean military has relied on a variety of time series techniques employing information from the equipment maintenance information system. However, alternative demand forecasting models must be investigated to improve accuracy. This study proposes a demand forecasting model that implements machine learning techniques to enhance the accuracy of spare parts demand forecasting, which is central to the military field. UAV spare consumption data were used to develop a classification model for predicting future demand. [Submitted: 17 November 2023; Accepted: 19 October 2024]