Research And Application Of WebGIS Cultivated Land Quality Prediction Model Based On Machine Learning Method

China is the largest agricultural producing country in the world,and its vast arable land resources and long history of farming culture make China top in the world in terms of agricultural strength.At present,the cultivated land resources in Our country are large in quantity and poor in quality.The evaluation of cultivated land quality grade carried out by local agricultural departments has clarified the next key point of the improvement and protection of cultivated land quality.In this paper,LUJANG County,Anhui Province was taken as the research area,and a prediction model of cultivated land quality grade in Anhui Province was established by the decision tree(DT)model and random forest(RF)model in machine learning algorithm.Design and develop the quality of cultivated land in anhui province rank query and prediction system,and combining the model embedded with Web GIS technology,in the form of GIS vector diagram,intuitively show the quality of cultivated land in anhui province level distribution,realize regional location and spot check,cultivated land quality related information query and analysis,the level of cultivated land quality forecast,and other functions,In order to provide reference for agricultural department.The main work and research results are as follows:(1)In view of the selection of influencing factors of cultivated land quality in Anhui Province,based on the national standard of cultivated land quality grade,the degree and importance of influencing factors of cultivated land quality were deeply considered,and the feature selection method was used to select the feature influencing factors.A total of 11 indicators were selected as the impact factors of this study,including the status of cultivated land use,soil type,surface texture,texture configuration,topographic position,irrigation capacity,drainage capacity,soil conventional nutrients(organic matter content,available phosphorus content,available potassium content)and soil p H value.(2)The index data of influencing factors of cultivated land quality characteristics are multi-dimensional,multi-type and multi-dimensional.After being processed by a series of methods such as principal component analysis(PCA),data Quantization and data Normalization,the data set has a strong applicability to the machine learning algorithm model.(3)The cultivated land quality grade prediction model based on DT and RF was established,and the accuracy of the model was evaluated by the evaluation indexes such as goodness of fit(R2)and root mean square error(RMSE).The evaluation results show that the cultivated land quality grade prediction models based on DT and RF have relatively stable prediction performance.Among them,RF has the best performance,the goodness of fit R2 for predicting cultivated land quality grade is 98.81%,RMSE is 0.27,and the model has high stability.(4)The cultivated land quality grade inquiry and prediction system of Anhui Province was designed and developed,and the system was tested back.System includes the cultivated land quality grade,soil nutrient,soil type query and visualization,cultivated land quality grade forecast simulation,and other functions,back to the testing results show that the model of cultivated extraction in 2020-2021 parts of cultivated land quality grade evaluation data has strong adaptability,test results and the real difference within the 0.1,the results of simulation result has reference significance.