The stability of a landside is affected by various non-linear factors. Consequently, a large amount of spatial data have to be analyzed to predict the stability of landsides. Geographic Information Systems (GIS)
have a distinct advantage in analyzing spatial data and they offer a good method for the prediction of landslides. Predicting landslides based on a neural network does not require a complicated mathematical model.
A neural network can be used for the prediction of landslides by studying a large number of samples. The method of landslide prediction in this paper is based a coupled Geographic Information System (GIS) and neural network. The approach analyses various factors that influence the stability of landslides. These factors include external factors (such as rainfall, earthquakes, anthropogenic activity) and internal factors (for example, rock structure and topography), which can be described quantitatively. These quantitative factors and their spatial distribution comprise the basic elements of the landslide database. The database of landside predictions
includes information on topography and hydrogeology. These quantitative factors are stored in the Geographic Information System and serve as the basis for prediction. To explain how to analyze the factors that affect
landsides, this paper uses the western mountain area of Hubei (a province of China) as an example. The advantages of using a neural network-based predication method are introduced in this paper. The basic theory of neural networks is also introduced in this paper, using the BP neural network as an example. The neural network acquires the relationship between the factors of the landside and its hazard index after a large number of samples have been used to train the neural network. The neural network can be then be used for the prediction of landsides.