Application Of Big Data Mining Method Based On Weka In Earthquake Precursor Data Procesing

In the environment of global information technology and Internet technology develop rapidly and the data information booming growth,The relevant scientific research of the earthquake also walk into this environment.As a consequence,big data and data mining technology in the research field of earthquake monitoring data aroused people's widespread attention.Big data by itself contains enormous potential value to promote the technology of data mining.That will be very meaningful and focused research to mine the earthquake precursor data which is large capacity,diversity,high-speed update and the potential value.Earthquake precursor data output greatly after digitally and networked transforming precursory network monitoring system Seismic stations in the period of national “fifteen”,So half of artificial and traditional data processing method cannot have satisfied the actual work requirements,Taking Taiyuan city earthquake precursor deformation data from 2011 to 2016 as the object of data mining,the basic idea of data mining method,the two main study content of this paper is following :One hand,Based on the open source software Weka data mining tools developed by JAVA to do precursor data pre-processing by settingup the Forecast environment mainly.According to the development trend of the time series data to build model for adding a amount of missing data.But the result data predicted data as the interpolation data.At last,this method has a promoting effect for the work of earthquake precursor data pretreatment.The other hand,According to the characteristics of earthquake precursor data test item,this paper adopted effective and convenience open Weka.And DFCM algorithm is embedded into Weka through MyEclipse.Then DFCM play to the important role in the interface of Weka,Analyzing each test item data by the fuzzy clustering analysis of measurement.Finally,concluded the following conclusions:Data preprocessing research includes the following work,firstly,according to the complicated and easily loss precursor data to deal with the missing data with the method of Weka-Forecast interpolation.Then to compare with the other industry interpolation results for the size of the mean standard error criterion.It shows that the Weka-Forecast interpolation applicability is much better than other common interpolation methods and the interpolation principle analyze precursory data is useful.Earthquake precursor data clustering analysis in Weka is successfully practical significance by changing the parameters of DFCM.Compared with half artificial detection standards,the accuracy of experimental result is good and the abnormal data was detected.This paper innovation point lies in big data mining technology is needed by the earthquake precursor data research in the development time.Mining the new knowledge could be congitived by clustering analysis data.This is a new innovative thinking model with Effective data mining tools.With that making this innovative thinking model to the real experiment,summarizes the new rule of precursory data.This is greatsignificance for other research problems in related precursory data.