運用資料採礦與3S技術建立坡地社區風險評估模式

Risk Assessment and Management by Using Data Mining and 3S Technique in Hillside Residential Communities

本研究採用3S（GPS 、GIS 、RS） 技術，評估台北縣汐止市坡地社區周緣環境地質災害風險度。風險模型以環境地質災害潛勢（ 或稱破壞機率）FP 與後果指數CS 之乘積函數構成。風險模式係利用馬氏法的判別分析法與貝氏定理求得環境地質災害潛勢，並以評分表量化損失因子為本研究之後果指數，加以結合計算風險值，此外採用統計方法中的集群分析法進行風險分級，區分為第一類、第二類與第三類優先防治社區，此為大尺度之風險評估。根據第一類優先防治社區挑選一處說明現調執行方式，依據建研所（2005） 現地調查危險度評分表量化調查點位風險等級，此屬小尺度之風險評估，最後結合兩種尺度風險評估成果繪製「坡地社區環境地質災害總體風險圖」，提供主管機關進行詳細調查、監測管理及災害防治策略研擬之依據，期能防患於未然，減少災害發生時之損失。

With the development of remote sensing (RS) and Geographical Information Systems (GIS) techniques, the environmental hazards in extensive areas can be evaluated in a short time. In this study, the risk model was developed, which is constituted by failure probability (FP) and consequence score (CS). The proposed risk model is based on discriminant analysis and Bayesian theorem to calculate FP. The CS is quantified in accordance with the grade table of consequence factor. FP and CS can be combined to calculate the risk value. In addition, statistical methods of cluster analysis are used to classify risk class, divided into high, medium and low risk communities. Based on risk analysis results, the selection of several risk communities is investigated to confirm the sensitive areas within the community. The combination of risk assessment results with the investigation results are drawn into risk maps to help relevant authorities to carry out community-based disaster prevention and countermeasures to develop reference.