The Analysis And Design Of Earthquake Prediction System Based On Extreme Value Theory

The earthquake disaster has brought many casualties and tremendous loss to human. Researchers all over the world are looking for a effective way to earthquake prediction. In recent decades, seismic studies have shown that in certain areas in the earthquake on a certain repetitive and cyclical. This offers a solution to the earthquake prediction research, which uses of mathematical statistical models to reflect a certain intensity of the earthquake cycle and earthquake probability earthquake, in order to achieve the purpose of earthquake prediction.Extreme value theory as a branch of mathematical statistics model, it has a wide range of applications like flood disaster, fire forecasting and financial risk analysis and any other fields. The prediction of earthquake have similar condition with them, the earthquake catalog which was already filtering the data use this model to statistical analysis, and combined with Poisson probability model can give a result of earthquake cycle and earthquake probability. This paper studies the extreme value theory in the application of earthquake prediction, and sets up a computer data analysis platform related to seismic data processing and analysis of mass, and support the research of seismic related areas use data to let it become more comprehensive.Study also found that the earthquake region has some correlation, it performed the association of certain geological structures relationship in the earthquake which called the arthquake-ralated phenomena. It is a common method that uses earthquake-related phenomena to studying earthquake prediction. Currently there are time series analysis of earthquake to find the relevant areas of the earthquake, also there are motheds using some mathematical models to analyze the earthquake catalog to find the correlation. But they have some limitations and obvious defects. This paper proposes a set of sequences based on angle matching algorithm related to the earthquake sequence described methods and similarity metrics. In this model, project the dimension be related of the earthquake sequence into a plane, and then processing the array linear by the feature points. The sequence of piecewise linear representation then transform to a sequence of point sets described, then using the definition of similarity metric model to dealing with a good sequence of calculations to be based on similarity is the largest in the database to find similar sequences above. This model overcomes the stability of the model using the pitch is poor, and the defect of isolated links between dimension.It provides a more efficient, stable and better overall performance algorithms & models, and provided a better technical support the earthquake prediction research. By the end, this paper uses a practical application to verify the effectiveness and correctness of the algorithm.