Immune Synergistic Methods For Earthquake Prediction

Earthquake is a kind of natural disaster which is the most harmful and serious to human life and property.If we can effectively predict the occurrence of the earthquake,do a good job in advance before the earthquake transfer work,will largely reduce the earthquake caused by the injury and property losses,conducive to the harmonious development of society.At present,the accuracy of earthquake prediction in the world is low.There are two main reasons: first,there are few precursory data that can be used for earthquake prediction.Second,earthquake prediction methods based on precursory signals are in the stage of analysis and processing.There are few methods of machine learning,and there are fewer training samples due to the low frequency of earthquakes.Inspired by the immune synergism of dendritic cells presenting antigens,t-cell recognition antigens,and dendritic cells and t-cells in the biological immune system,a seismic prediction model based on immune synergism was established in this paper.Improve and improve prediction accuracy from two aspects: 1)generate TC detector set through negative selection.Since the negative selection does not need abnormal samples to generate detector,it can effectively overcome the problem of too few samples in earthquake occurrence and reduce the dependence on abnormal samples.2)update the TC detector set through the DC antigen presentation process,and delete the detector with high affinity for normal antigen,which can effectively reduce the false alarm rate.The main work is as follows:Firstly,the importance and existing problems of earthquake prediction are put forward,and the difficulty and predictability of earthquake prediction are explained.The research based on earthquake precursors and historical seismic data is introduced.The electromagnetic and geoacoustic data collected in the AETA multi-component seismic monitoring system are introduced,and the good seismic reflection effect is presented.Secondly,the research status of biological immunity and computer immunity is introduced.By analyzing the typical models of biological immune mechanism and computer immune system,and combining the problems existing in earthquake prediction with the goals to be achieved in this paper,it is shown that the earthquake prediction method with immune mechanism can improve the accuracy of earthquake prediction.Then,the earthquake prediction model of immune coordination mechanism is proposed.The model includes feature index extraction,data preprocessing,initial TC detector set generation,DC signal acquisition and state detection,matching and result classification.The initial TC detector set was generated through negative selection with training data,and the DC antigen presentation state was detected with test data.The presented antigen was matched with the initial TC detector set to obtain the matching result.The TC detector set was updated according to the presented DC state,and the matching results were corresponded to whether the earthquake occurred or not.Finally,take the data collected in the "AETA multi-component seismic monitoring system" as the source,select the data within the specified time period to extract the characteristic index,and then conduct dimensionality reduction processing on the characteristic index to form the final data set through PCA.Then,the earthquake prediction method of immune mechanism proposed in this paper was experimentally verified.The neural network earthquake prediction method was used as a comparative experiment,and four statistical indicators proved that the method proposed in this paper had better prediction effect.