Resumo

GARCIA RODRIGUEZ, Frank Gustavo  e  HERNANDEZ SANTANA, José Ramón. PREDICTIBILITY OF GEOMORPHOLOGICAL RISKS IN THE NECAXA HY-DROELECTRIC SYSTEM, SIERRA NORTE DE PUEBLA, MEXICO.. Ser. correl. geol. [online]. 2017, vol.33, n.2, pp.1-10. ISSN 1666-9479.

This paper tries to state how landslides can be treated as objects in a contextual analysis. The goal of this research project is to provide a framework that integrates some existing models. Multi scale segmentation and Object-oriented analysis (OOA) have shown good results in the landslide inventory. The combination of the above methods and data mining techniques have allowed for hazard assessment procedures based on susceptibility maps to improve predictions rating landslide threats in the middle portion of the Necaxa River basin. The results were contrasted by the models of frequency ratio, entropy ratio, classification and regression trees and multivariate adaptive regression splines. The models performance were evaluated using the area under the receiver operating characteristic curve (AUC), at time statistic were used to compare the spatial heterogeneity of the typological units. The maximum and minimum probabilities of quality achieved, 0.87246 and 0.728410 respectively, allow us to conclude the validity of the methodology used.

Palavras-chave : Landslide hazard. Data mining. Necaxa. Mexico..

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )