| After years of marine geological survey,China has accumulated abundant marine geological survey data.Because marine geological survey and research involve many disciplines such as marine geology,geophysics,geochemistry,surveying and mapping science,geographic information science,researchers are difficult to fully discover and use the hidden information behind these data.Data mining technology can integrate and learn large data sets and generate useful patterns and rules,which are helpful to discover the laws of marine geology.This paper makes full use of data mining,computer and other technical means to develop a marine geological data mining system,and takes the petroleum and gas resources evaluation of a certain area in Laoshan Uplift of the South Yellow Sea as an example.The system identifies the areas with the largest,larger,smaller and smallest petroleum and gas potential,and determines the petroleum and gas enrichment exploration blocks in the study area.It is proved that the system designed and developed in this paper can meet the requirements of marine geological data mining,and has strong practicability,which provides new ideas and new tools for the research of marine geological data mining.The main work of this paper is summarized as follows:(1)This paper fully investigates the development and application of data mining technology and marine geological survey data.Based on the systematic study of marine geological data preprocessing,data analysis and comprehensive evaluation methods,the marine geological data mining system is designed.This marine geological data mining system is developed in the parameter interaction between the pyinstaller Library of Python and process of windows.The system includes mainstream clustering,regression,classification and other data mining methods.Users can use data standardization,preprocessing,principal component analysis,K-means clustering,random forest classification and other related data mining methods to realize the mining,analysis and evaluation of complex marine geological data information.When the system is used,there is no need to configure the environment,which greatly simplifies the work steps and improves the work efficiency.(2)This paper takes a region in Laoshan Uplift of the South Yellow Sea as an example,according to the steps of data preprocessing,evaluation index extraction and weight setting,petroleum and gas resources evaluation and result analysis,the practicability of the system is tested.The research area has not been drilled yet,so there is no label information.In this paper,seven clustering methods,such as k-means clustering,self-organizing mapping clustering and grey correlation method from the marine geological data mining system,are used to comprehensive evaluation.By comparing the comprehensive evaluation results of these eight methods with the well location to be drilled in the research area,this paper concludes that the data mining methods suitable for the research area are k-means clustering,aggregation clustering and self-organizing mapping clustering,and also proves the reliability of the results;The proportion rule of petroleum-gas potential area obtained by the above three methods is the same,which further reflects the scientificity,rationality and reliability of the comprehensive evaluation results. |
