Establishment of Predicting Landslide susceptibility for Alisan Forestry Railway by Logistic Regression Model

This study aimed to predict landslide susceptibility by using Logistic regression, a multivariate analysis method, and examined the performance of this approach. In order to select the factors efficiently, the calibration and selection procedure were performed. As a result, a suitable combination of the factors for establishing landslide susceptibility model was presented. The evaluation procedure of the susceptibility value of the factors was also presented. The classification error matrix was used to evaluate the accuracy of landslide predicted by the present model. Furthermore, the map of potential landslide was delineated. A landslide susceptibility map with susceptibility index was proposed for engineering and disaster prevention consideration.本研究以羅吉斯迴歸法建立阿里山森林鐵路附近邊坡之山崩潛感預測模式，為有效挑選模式內之山崩潛感因子，利用山崩與非山崩次數分佈圖、崩壞比圖及檢測資料是否符合常態分配後進行因子篩選，針對各山崩潛感因子的山崩組與非山崩組資料進行取樣後，以羅吉斯迴歸法計算出各因子的羅吉斯迴歸係數，並建立山崩潛感模式。並為了解所建立模式之優劣，利用分類誤差矩陣表對建置完成之山崩潛感模式進行結果驗證，最後將全區篩選過後之因子資料套用至此羅吉斯迴歸模式，計算得到各網格之潛感值並且繪製出山崩潛感圖，相關成果可提供日後阿里山森林鐵路附近邊坡管理及山崩預警之參考