کد مقاله کد نشریه سال انتشار مقاله انگلیسی نسخه تمام متن
5780745 1635355 2017 11 صفحه PDF دانلود رایگان
عنوان انگلیسی مقاله ISI
Landslide susceptibility mapping & prediction using Support Vector Machine for Mandakini River Basin, Garhwal Himalaya, India
ترجمه فارسی عنوان
نقشه برداری و پیش بینی قابلیت حساسیت به زمین لغزش با استفاده از ماشین بردار پشتیبانی برای حوضه رودخانه منداکینی، هارمولای هارمون، هند
موضوعات مرتبط
مهندسی و علوم پایه علوم زمین و سیارات فرآیندهای سطح زمین
چکیده انگلیسی

In recent years, various machine learning techniques have been applied for landslide susceptibility mapping. In this study, three different variants of support vector machine viz., SVM, Proximal Support Vector Machine (PSVM) and L2-Support Vector Machine - Modified Finite Newton (L2-SVM-MFN) have been applied on the Mandakini River Basin in Uttarakhand, India to carry out the landslide susceptibility mapping. Eight thematic layers such as elevation, slope, aspect, drainages, geology/lithology, buffer of thrusts/faults, buffer of streams and soil along with the past landslide data were mapped in GIS environment and used for landslide susceptibility mapping in MATLAB. The study area covering 1625 km2 has merely 0.11% of area under landslides. There are 2009 pixels for past landslides out of which 50% (1000) landslides were considered as training set while remaining 50% as testing set. The performance of these techniques has been evaluated and the computational results show that L2-SVM-MFN obtains higher prediction values (0.829) of receiver operating characteristic curve (AUC-area under the curve) as compared to 0.807 for PSVM model and 0.79 for SVM. The results obtained from L2-SVM-MFN model are found to be superior than other SVM prediction models and suggest the usefulness of this technique to problem of landslide susceptibility mapping where training data is very less. However, these techniques can be used for satisfactory determination of susceptible zones with these inputs.

ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Geomorphology - Volume 295, 15 October 2017, Pages 115-125
نویسندگان
, , , ,