Research On The Evaluation Model Of Arable Land Quality Based On IPSO-LSTM

As a country with a large population,China strictly adheres to the red line of 1.8 billion mu of arable land to ensure food production and safety,which is related to my country’s economic and population development,and the quality of arable land directly affects the harvest and health of grain.Evaluation research is imminent.In recent years,the continuous development of machine learning methods has made it possible to achieve breakthroughs in applications in important fields such as agriculture,medical care,transportation,finance,and manufacturing.The traditional methods of evaluating the quality of cultivated land are inefficient.The use of machine learning methods can not only improve cultivated land The accuracy and efficiency of quality grade evaluation can also effectively adjust the agricultural industry structure and rationally plan and utilize cultivated land,avoid excessive fluctuations in the output level of cultivated land and the quality of agricultural products,and stabilize my country’s grain market.The research results and contributions of this paper are as follows:(1)The research and analysis of the cultivated land quality evaluation factor data.Through the preliminary analysis of the cultivated land quality evaluation factor data,the outliers,missing values and other information are sorted out,and the sorted data are preprocessed by using data processing tools,and then the covariance,covariance matrix and Pearson correlation coefficient are used to filter the data.The important cultivated land quality evaluation indicators are used for feature construction,and the evaluation of cultivated land quality grades can be effectively converted into supervised classification research.(2)A LSTM-based arable land quality grade evaluation model is proposed.This model can fully excavate the hidden laws behind the data of cultivated land quality evaluation factors,and conduct cultivated land quality grade evaluation.In addition,the cultivated land quality grade evaluation models of MLR,RF and DT are established,and unified model evaluation indicators are designed.The results of the LSTM model are the best.Therefore,we choose to further study the LSTM-based cultivated land quality grade evaluation model.(3)The LSTM-based cultivated land quality grade evaluation model is optimized.The IPSO algorithm was used to optimize the LSTM evaluation model,and it was optimized into the IPSO-LSTM cultivated land quality grade evaluation model.In this study,the parameter optimization method was introduced.By improving the particle swarm algorithm(PSO),the improved particle swarm algorithm(IPSO)was used to optimize the parameters of the LSTM evaluation model.The results of the model optimization were further improved,and the training set accuracy was 96.8%,the accuracy of the test set is96.3%,the accuracy of the validation set is 97.2%,the recall rate of the validation set is97.1%,and the accuracy of the validation set with an allowable error of 1 level is 99.9%.Effect.(4)Developed the Chaohu Cultivated Land Quality Grade Evaluation System.After completing the above model research,the research model is transformed into a web interface of graphical operation,developed based on the Vue framework,the data input part of the model is realized as an operable interface,and the data is transmitted by calling the API interface.Calculate the evaluation model,and then return the results.The results are displayed graphically in Echarts.Finally,the development of the system is completed through actual operation and verification.