중동부유럽
연구정보
[IT] Landslide susceptibility in a hilly region of Romania using artificial intelligence and bivariate statistics
루마니아 국외연구자료 연구보고서 Applications of AI in Mining and Geotechnical Engineering 발간일 : 2024-01-12 등록일 : 2024-04-04 원문링크
자료이용안내
국내외 주요 기관에서 발표하는 자료들을 수집하여 제공하고 있습니다. 수록 자료의 자세한 내용은 해당 기관으로 문의하시기 바랍니다.
In the current study, a landslide susceptibility estimation procedure was presented using two models. The first model is represented by the frequency ratio method, while the second model was generated by combining frequency ratio with multilayer perceptron. The present research paper focused on the area of the upper and middle basin of the Cricovul Sarat river in Romania. A number of 321 locations of landslides and 10 landslide predictors were used to determine susceptibility. It should be mentioned that the ROC curve was involved as validation algorithm for the results. This indicated that both models used had a high accuracy, area under curve (AUC) having values above 0.85. At the same time, the most exposed surfaces to landslides were those with slope values over 25°.
본 페이지에 등재된 자료는 운영기관(KIEP) 및 EMERiCs의 공식적인 입장을 대변하고 있지 않습니다.
