Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/88370
標題: 應用羅吉斯迴歸建立台21線公路崩塌潛感預測模式
Landslide Susceptibility Prediction Model for Route No.21 Using Logistic Regression Analysis
作者: Yi-Wei Chao
Kuang-Tsung Chang
趙逸幃
張光宗
關鍵字: Route No.21
landslide
logistic regression
台21線
崩塌
羅吉斯迴歸
摘要: This study illustrates a landslide susceptibility prediction model for Route No.21 using logistic regression. In this model, possible factors of landslide are selected and can be classified into three categories, topography, geology, and hydrology. We built this model based on the data of past four typhoons, Herb, Sinlaku, Mindulle, and Toraji. Then we calculate the logistic coefficient of each factors and verify this model by the data of typhoon Morakot. The precision of this model is verified by classification error matrix and ROC curve. At last, we put the data of four typhoons into this model to calculate the susceptibility of each slope unit, establishing landslide model along the slope unit of route No.21 in Chenyoulan creek watershed. The accuracy of this model is 75.2%. It is suitable for the landslide prediction and precaution in this area in the future.
本研究利用羅吉斯迴歸建立台21線陳有蘭溪沿線邊坡單元之崩塌潛感模式,對可能造成崩塌之因子進行篩選,崩塌因子可分成三大類型:地形、地質與水文。將歷年四次颱風(賀伯颱風、桃芝颱風、敏督利颱風及辛樂克颱風)資料建立模型,並計算各因子的羅吉斯迴歸係數,最後再以莫拉克颱風驗證。驗證結果以分類誤差矩陣與ROC曲線判斷其模型之準確度。最後以四次颱風資料帶入模型內計算各邊坡單元之潛感值,建置陳有蘭溪流域公路沿線邊坡單元崩塌模型,模型準確度為75.2%,可供日後崩塌預測、預警性封路之參考。
URI: http://hdl.handle.net/11455/88370
Appears in Collections:第46卷 第04期

文件中的檔案:

取得全文請前往華藝線上圖書館



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.