Research And Implementation Of Grain Production Forecasting Model Based On WOFOST Model And Support Vector Machine

Agriculture is the mother industry of human beings.Whether in less developed developing countries or in developed countries where science and technology are both economically developed,What the government and people of the country care most about and attach importance to is the development of agriculture.Food is the foundation of social progress and the lifeblood of economic development.It is related to the prosperity of the country,the people's lives and social stability.With the continuous development of science and technology,the country's strength continues to increase.Traditional farming has gradually been replaced by mechanized agriculture.The once thinking of relying on heaven for food has gradually turned into a new scientific agriculture,and China's agricultural modernization has ushered in a new era.The prediction of food production has also changed from past empirical predictions to current scientific inferences.Accurate food production forecast can help the government to adjust the programs and strategies in grain production,which will help improve the efficiency of food production and greatly accelerate the development of new scientific agriculture.Therefore,experts and scholars at home and abroad have focused their research and attention on how to reasonably and accurately forecast food production in order to obtain macroscopic control over the future development trend of grain production.Support Vector Machine(SVM)is a learning method developed based on statistical theory.It has a good performance in classification and regression problems.In particular,it uses the principle of structure minimization to solve small sample problems,and It has been highly recognized by experts and scholars all over the world.Therefore,it can solve the small-sample regression problem of food production forecast.The WOFOST crop growth simulation model is a quantitative analysis model of annual crop growth and yield.After numerous domestic and foreign experts and scholars have studied,it has been confirmed that it has been universally verified and applied on a global scale.The model has a good simulation effect on the rhizome crops such as corn,wheat,and rice,and the application field is becoming more and more extensive.This paper firstly introduced the background and significance of grain production forecasting,and expounded the research status of WOFOST crop growth model and support vector machine.Then explained the principle of WOFOST crop growth model and the theoretical basis of support vector machine in detail.Further,based on some existing researches,the SVR-WOFOST model is constructed by combining the SVM regression method with the WOFOST model.The WOFOST model was used to analyze the meteorological factors,soil factors and the nature of the crops,and regression prediction was carried out by support vector regression.In the selection of impact factors,the use of grey correlation analysis as a reliable evidence of the selected impact factor.In the end,Changchun City was selected as the experimental area,and corn yield was used as the experimental subject for empirical experiments.This paper presents a new food production forecast model SVR-WOFOST based on WOFOST crop growth model and support vector machine algorithm.This model not only applies the method of support vector machine regression,but also uses the characteristics of the WOFOST model.Based on the standard support vector machine prediction model,it adds natural influence factors to improve the accuracy of the model.In practical applications,it is possible to combine the daily weather data for the next three months obtained from climate predictions,and to forecast the grain production of the current year in the middle of the year,which is more in line with practical applications and more conducive to promotion.Especially for forecasting in extreme weather conditions is more accurate and reliable.Through empirical experiments on the yield of corn in Changchun City,the accuracy of the model and the standard SVM model is compared.The experimental results show that the support vector machine combined with the WOFOST model has a good performance in predicting corn yield,especially in the disaster year relative to the standard support vector machine.