assessment land susceptibility to gully erosion using ensemble data-miningTechniques (Case study: Shahroud Basin)

at first, a gully erosion inventory map was prepared using extensive field surveys and interpretation of aerial photographs, of which 172 gully erosion, 70% (121 gully) are used for modeling and 30% (51 gully) are used for validation purposes. In the next step, for the selection of parameters, after their initial identification, the multicollinearity analysis test was performed using coefficients of tolerance and variance inflation factor and the parameters with multicollinearity were deleted due to the reduced accuracy of the modeling, , Finally, 12 parameters were selected for modeling. The results of determining the significance of the criteria by entropy index method showed that elevation, lithology and NDVI parameters had the greatest effect on the occurrence of gully. In order to validate the model, the prediction rate and success rate as well as the SCAI index were used. The validation results showed that the combined model with a prediction rate of 956.0 (95.6%) and a success rate of 92.33 (92.3%) had excellent predictive accuracy and compared with the entropy index and evidential belief function with prediction rates of 0.932 and 0.917, and the success rates of 0.911 and 0.901, have higher accuracy. According to the results of the SCAI, class differentiation was appropriate in the combined model. According to the results, 28.95 percentage of the study area is located in high and very high susceptibility classes. The results of this research can be used by landuse planners to expansion development activities.