The application of the intelligent algorithm in the prevention and early warning of mountain mass landslide disaster

Abstract

The objective is to realize the early warning of disasters caused by mountain mass landslide. The probable causes of mountain mass landslide were predicted and analyzed in prior by constructing the back propagation (BP) neural network algorithm, which included earthquake, rainstorm, human activity, landslide displacement, slope gradient, and soil texture. In terms of the early warning of mountain mass landslide caused by earthquakes, rainstorm, or human activities, the accuracy of BP neural network algorithm was relatively high; especially, given the relevant data such as the free slope gradient, slide surface slope gradient, and soil texture of the mountain mass, the accuracy of BP neural network algorithm in the early warning of mountain mass landslide could reach 94.7%; the proposed algorithm could predict not only the occurrence possibility of mountain mass landslide but also the severity and possible range of mountain mass landslide. In addition, in terms of stability, the standard deviation of the proposed algorithm was 0.0062, which indicated the fairly good stability of the algorithm. The possible mountain mass landslide was predicted based on the comprehensive analysis of probable causes of mountain mass landslide and their proportions of weights by using the BP neural network algorithm; the accuracy and stability of the proposed algorithm were excellent, which also showed that the occurrence of mountain mass landslide was a comprehensive consequence caused by various factors and could never be predicted through singular factor analysis. Thus, the understanding of both BP neural network algorithm and mountain mass landslide was greatly improved.