Landslide susceptibility assessment is essential for development activities and disaster management in the mountainous regions to identify the landslide-prone areas. The present study aimed to evaluate and compare the efficacy of data driven quantitative models of landslide susceptibility assessment using geospatial tools in Jhelum valley of the Himalayas. This area suffers from extreme rainfall events due to the local climate and has experienced significant and widespread landslide events in recent years. Four probabilistic data-driven models are employed for this purpose, which includes the weight of evidence (WOE), information value method (IVM), frequency ratio (FR), and certainty factor (CF). These assessed models are based on integrating landslide contributing factors and a ground truthing-based landslide inventory of 437 landslides. The landslide susceptibility maps were presented by categorizing the study area into very low to very high susceptibility zone by Jenks natural breaks method. The performance of models was evaluated by a sensitivity analysis using Receiver Operator Curve (ROC) method. The ROCvalidated results of success rate curves for WOE, IVM, FR and CF were 80 %, 78 %, 77 %, and 76 % respectively. The prediction rate curve of WOE, IVM, FR, and CF was 78 %, 77 %, 75 %, and 78 % respectively. The results showed the reasonable efficiency of applied models for landslide susceptibility assessment in the study area and applicable to regions with similar geomorphological conditions. Conclusively, the comparison of applied models revealed the promising results of used approaches., Salman Farooq and Mian Sohail Akram., and Obsahuje bibliografii
Artificial neural networks (ANNs) have been used to construct empirical nonlinear models of process data. Because networks are not based on the physical theory and contain nonlinearities, their predictions are suspect when extrapolating beyond the range of original training data. Standard networks give no indication of possible errors due to extrapolation. This paper describes a sequential supervised learning scheme for the recently formalized Growing Multi-experts Network (GMN). It is shown that the Certainty Factor can be generated by the GMN that can be taken as an extrapolation detector for the GMN. The On-line GMN identification algorithm is presented and its performance is evaluated. The capability of the GMN to extrapolate is also indicated. Four benchmark experiments are dealt with to demonstrate the effectiveness and utility of the GMN as a universal function approximator.