Study On Data Processing Of Soil Attribute Based On Data Mining Technology

The traditional data processing of soil attribute is generally based on the specific application oriented,which processing of a single industry and small scale data and ignoring the mathematical characteristics of the data,and then problems appear,such as tedious data processing,professional analysis,not sharing of cross industry data,etc.Data mining is the data processing technology based on the mathematical characteristics of the data itself,ignoring the professional significance of the data to explore and discover its intrinsic value.This study systematically introduced the data mining technology of soil attribute data processing,and its processing method and traditional soil science methods are compared to explore their relevance,can open a large-scale data processing and cross industry data sharing channel for traditional soil data processing,and provide a theoretical basis for the improvement of soil science data processing method and in line with the era of big data.The main purpose of this study is to explore the relationship between the two methods by comparing data mining technology and traditional soil data processing methods.Based on the data of 176 soil profiles in 74 counties of Guangdong Province,this paper studies the application of data mining in the data processing of soil properties andthe results are as follows:(1)The data mining analysis and processing of soil attribute data model are constructed,including clustering,ID3 decision tree method,grey correlation method,data mining visualization tools etc.in this paper.Using data mining technology to clean out,soil property metadata sorting,conversion,will be deleted or modified w duplicate data,empty data,obviously unreasonable data,deviated data,dirty data,and ultimately the 797 final data as the research object are determined.Closely relationship exists between the soil natural attribute itself the characteristics of the data based on the data mining classification method and based on the traditional classification and soil science disciplines(evaluation)method through the study.(2)There is a close correlation between the order classification of soil classificationand the classification of data mining methods based on the data characteristics of soil attribute data.In the rough or non professional classification and cross industry data applications can consider the application of soil order classification based on data characteristic as the reference category of soil classification,andnon soil field of massive soil attribute data to facilitate the use of data mining technology in the field of the analysis results.(3)Data mining method is used to modify and validate the model.the rough classification and evaluation of paddy soils in Guangdong province were carried out based on the data collected from various cities in Guangdong Province.The author finds out the deficiency of the original evaluation method,and puts forward a new comprehensive index method in the process of evaluation.By classification and evaluation method for the comprehensive index score,the results are as follows: first class accounts for 6.78%,two class accounts for37.14%,etc.,threeclass accounts for 55.96%,and four class accounts for0.13%.(4)Using the ID3 algorithm of data mining technology to analyze the factors that influence the evaluation of soil,and construct the searching model of the most influential factor.According to the size of the influencing factors: organic matter > total nitrogen >available phosphorus > alkali nitrogen > available potassium >pH.Through this method,the influence factors of soil quality can be quickly sorted and the quantitative indexes can be calculated.(5)The correlation analysis model of soil attribute data is constructed and studied.Correlation analysis was performed on the sampled soil attribute data by grey correlation method and relevance is ranking according to the calculation result of correlation factor.The correlation coefficients of 17 kinds of soil attribute data respectively and soil organic matter content calculation,sequencing results: total nitrogen,total phosphorus,total potassium,alkali nitrogen,availablephosphorus,availablepotassium,cation exchange capacity,exchangeable potassium,exchangeable sodium,exchangeable calcium,exchangeable magnesium,exchangeable salt,base saturation,total iron and free iron,amorphous iron,iron free degree.(6)The results of soil data analysis were visualized and visualized by bar graph,line graph,pie chart,scatter plot,and box plot,using tableau visualization technology of the data mining tool.