隨著當代資訊科技的進展，我們保存了越來越大量的資料，但隨著資料龐雜所衍生的雜訊也越來越多，因此也造成資料探勘品質的下降。本研究利用地理資訊系統的空間資料處理能力，以119急難救護的空間資料為操作案例，針對花蓮縣民國96年6月到97年12月間，合計24,108個119急難救護資料，提出如何在資料探勘技術下，將這些過去不易保存或深入探索的資料，透過地址對位，並從村里的大比例尺地理環境，分析案例發生區位、人口密度、道路密度、消防分隊責任範圍，利用地理加權迴歸，檢視整個花蓮縣急難救護的醫療資源，並推估村里間醫療資源利用的空間分佈，藉以發掘出發現先前關心卻未曾有效擷取的重要知識，進而擬定策略或支援決策，提出如何在合理的醫療資源分配原則下，建議巡迴醫療的指派方式。

At present, databases are highly susceptible to noisy data, and due to their immense size, lead to low-quality mining results. This research proposed a GIS preprocessed method to improve the efficiency and ease of data mining processes. In the case studies, we used 119 emergency care data to locate the addresses of 24,108 pieces of 119 emergency care information, analyzed the regions of case occurrences, population density, road network density, and areas of responsibility of fire brigades through the proportional scales of geographical environments of villages and townships. This study then used geographically weighted regression (GWR) to inspect the medical resources of emergency care, regarding which remote townships should first be enhanced for insufficient equipments and manpower by utilizingmobile medical services. Using GIS digitalized technologies, researchers were able to acquire important knowledge of concern, but were unable to extract from among increasingly complicated primitive data in order to further develop strategies and support policies.