Anomaly Recognition of Ultra Low Frequency Electric Data Based on Artificial Neural Network

In order to search the abnormal data of seismic satellite quickly, we extract the ultra low frequency electric field waveform data of 10 days before the Wenchuan earthquake in this paper. The back propagation neural network classification model is designed and the self-organizing feature map network clustering model is used to verify back propagation network using the mean, variance, skewness and kurtosis as the feature information. Results show that anomaly areas are located in the larger area of the south of the epicenter of the Wenchuan earthquake, which is perhaps due to the powerful seismic wave energy of south of Wenchuan causing the strong spatial ionospheric disturbance; the correct recognition rate of the normal class of back propagation net classifier reaches 98.13%, the recognition rate of the exception class of that is 96.75%; The clustering results of self-organizing feature map network and the back propagation network classifier are practically in good agreement.