A predicton model for the seismic intensity in meizoseismal area based on Genetic Neural Network

In this paper, a novel model for seismic intensity prediction in meizoseismal area is proposed based on BP neural network. 322 earthquake events of MS 5.0 or more occurred in the mainland of China from 1966 to 2017 with detailed intensity records. From the collected sata set, seven factors related to the seismic intensity are selected. Principal component analysis is performed on these factors to generate the network input, while the intensity of the epicenter is the output of the network. The BP neural network is optimized by using genetic algorithm. Simulation results show that, compared to the traditonal prediction methods, the proposed prediction model based on BP neural network has obvious advantages in terms of prediction accuracy and correct prediction probability.