Integration of Information Theory, K-Means Cluster Analysis and the Logistic Regression Model for Landslide Susceptibility Mapping in the Three Gorges Area, China

摘要: In this work, an effective framework for landslide susceptibility mapping (LSM) is presented by integrating information theory, K-means cluster analysis and statistical models. general, landslides are triggered many causative factors at a local scale, the impact of these closely related to geographic locations spatial neighborhoods. Based on facts, main idea research group study area into several clusters ensure that in each affected same set selected factors. idea, proposed predictive method constructed accurate LSM regional scale applying model area. Specifically, factor first classified natural breaks with optimal number classes, which determined adopting Shannon’s entropy index. Then, certainty (CF) class estimated. The selection based CF values factor. Furthermore, logistic regression used as example models using prediction. Finally, global map obtained combining maps. Experimental results both qualitative quantitative indicated can achieve more maps when compared some existing methods, e.g., overall prediction accuracy 91.76%, 7.63–11.5% higher than those methods. Therefore, technique very promising further improvement