鑑於全球對大眾捷運系統（MRT）之依賴，優化後之維護規畫成為實現高效能運營成本之關鍵；然而，確定影響捷運維護成本等重要因素之複雜性，仍為持續存在之挑戰。本研究運用約略集增強K-近鄰演算法（KNN）調查6項捷運維護成本中之重要影響因素：對捷運之需求、維護團隊執行檢修任務所需之差旅（OV）需求、員工薪資、工地成本、捷運直接費率、以及用於載運維護人員之公務車輛直接費率。 本研究採用之研究方法包括文獻回顧、數據收集、以及用於確定最具影響力因素之約略集增強分析。研究結果顯示，捷運直接費率為核心因素，頻率為78.95%，共有820次獨特出現。其次為OV直接費率，頻率為76.48%，共出現736次。薪資和需求（MRT和OV）等其他因素之頻率約為60%（獨特出現325-349次），而工地成本之頻率為55.66%，共出現193次。本研究提供大眾捷運系統（MRT）之利害關係人一量化方法，使其得以更準確地估計維護成本。此種改良方法可提升決策過程之成效，尤其益於預算規劃及資源分配具有益處。 ;The global reliance on Mass Rapid Transit (MRT) systems necessitates optimized maintenance planning for efficient operational costs. The complexity of identifying significant factors affecting MRT maintenance costs has presented persistent challenges. This research uses the Rough Set Enhanced K-Nearest Neighbor (KNN) method to investigate the factors in MRT maintenance costs. The study targets six influential factors: demand MRT, demand OV, staff salary, site cost, MRT direct rate and official bus direct rate for transportation of the crew maintenance. The research methodology implemented in this study comprised a structured blend of literature review, data collection, and Rough Set Enhanced analysis used to determine the most influential factor. The MRT direct rate, with a frequency of 78.95% and 820 unique occurrences, is the core factor. The OV direct rate follows closely at 76.48% frequency and 736 occurrences. Other factors like Salary and Demand (for MRT and OV) show frequencies around 60% (unique occurrences 325-349), while site cost registers a frequency of 55.66% with 193 occurrences. The study provides a quantifiable method for MRT stakeholders to estimate maintenance costs more accurately. This improved approach enhances decision-making processes, which is particularly beneficial for budget planning and resource allocation.
