Study On Soil Salinization Data Mining Based On Fuzzy Algorithm

Soil salinization refers to the comprehensive effect of specific climate,geology,soil texture and man-made irrigation and other improper factors,land quality degradation caused by,generally in arid and semi arid region.Salinization,as one of the forms of land desertification,is mainly distributed in arid and semi-arid areas of Western China.It is one of the major obstacles to the development of agriculture and the stability of oasis ecological environment in Western china.Xinjiang is a typical arid area,and soil salinization is widespread.A large number of saline soil caused by the decline in soil productivity,soil imbalance,not only affects the strategic position of Xinjiang in China's production base of grain and cotton,and hinder the sustainable development of oasis economy and ecological security in arid area.Therefore,in Xinjiang for the prevention and improvement of saline as the goal,extraction and grasp the dynamic changes of soil salinization of information is a work of practical significance,which can provide important basis for decision making on land management,agricultural production and regional planning and other related government departments.With the rapid development of information technology,the amount of data accumulated by human beings is increasing.For these large amounts of data,human beings are not satisfied with the use of traditional queries and statistical analysis to find deeper sub laws.Therefore,in recent years,a new technology--data mining technology has emerged.Data mining refers to the process of discovering,hiding,and using valuable information and knowledge hidden from a large amount of incomplete,vague data.Fuzzy phenomena can be seen everywhere in the field of social science or natural science.It is difficult to solve these fuzzy concepts by using traditional methods,and the fuzzy algorithm solves this difficult problem.In recent years,fuzzy algorithm has been widely used in the field of data mining,and has achieved great application value.Soil salinization exists fuzzy type definition,this paper solve the fuzziness of soil salinization congenital through the fuzzy algorithm,and fuzzy algorithms are one of the commonly used data mining,the main task of data mining prediction,Risk assessment and classification.Therefore,this paper studies the soil salinization from the main tasks of data mining.The main research work of this paper is as follows:(1)the birth and development of fuzzy algorithm and the difference between it and the classical set theory are introduced.Then,the basic theories are introduced.In depth understanding of the basic concepts of fuzzy algorithms and the use of methods for subsequent fuzzy algorithm to lay a good foundation.(2)according to the fuzzy algorithm in data mining applications are studied.Onthe basis of field investigation and laboratory experiment data of observation station data on the effects of kinds of factors affecting soil salinization,each of these data should be used for several commonly used fuzzy theory method in data mining,i.e.prediction method to analyze the case of fuzzy neural network.The prediction of soil salinity by fuzzy neural network shows that the measured values of most sampling points coincide with the predicted values of the fuzzy neural network.Some of the prediction results of some points have some deviations,thus affecting the overall prediction effect.The relative error is basically less than 1%,the average error is0.2082%,the minimum error is 0.0012%,and the maximum relative error is 0.9108%.The correlation coefficient between the measured value and the predicted value is0.536,which shows that the prediction ability is good,And risk evaluation of soil salinization in the study area based on model outputs.(3)remote sensing image classification is one of the most important tasks of data mining,the large amount of information with remote sensing images,and is currently in the salinization monitoring by remote sensing,remote sensing image classification method is based on traditional visual interpretation or the introduction of auxiliary amount of main salinization information extraction.Due to the presence of congenital salinization type definition itself overlap problem caused by different types of spectral inhomogeneity of aliasing and the same type of spectra,classification of the boundaries is often not clear,with great ambiguity;fuzzy classification is widely used in remote sensing information processing,can realize the salinization information extraction and classification with high accuracy.This paper using the fuzzy C-mean clustering method and fuzzy classification method based on object oriented information extraction of soil salinization of the remote sensing image,and the two methods were compared through accuracy evaluation.The results showed that the independent component analysis and classification method combined with fuzzy Cmeans algorithm has the best performance and the highest classification accuracy for saline soil classification,the overall classification accuracy of 82.47%,Kappa coefficient is 0.742,compared to the single classification of fuzzy C-means increased by 4.94% and 0.032,K-increased by 9.8% and the average clustering method 0.080,significantly improves the classification accuracy.Finally,the fuzzy classification based on object oriented analysis,found using object oriented method of fuzzy algorithm in soil salinity remote sensing information extraction to achieve high accuracy of classification results,classification accuracy reached 86.25%,kappa coefficient is 0.826,fully reflects the high one image in the terrain details the advantages and application value.