Research On Prediction Method Of Grain Yield In AnHui Province Based On Machine Learning

Grain yield prediction is the key to grain storage,farmland management and national agricultural decision-making,in addition,it also an important part of national food security assessment and food policy formulation,crop growth and yield are the key to adjust agricultural planting system and agricultural operation and management.a series of factors,including population increase,decrease of arable land and water resources,environmental degradation and global climate change,have a significant impact on agricultural production and threaten food security.it is estimated that by 2050,the total population of the world will reach 9.15 billion,which will have an important impact on world agriculture.food security guarantee is crucial.therefore,accurate regional crop growth monitoring and yield prediction are crucial for guiding agricultural production,national food security and sustainable agricultural development.In this paper,after having an deep understanding on the geographical condition and the trend of grain production in Anhui province,first of all,the influence mechanism of various factors on grain yield in Anhui province is elaborated in detail,and this study determines the dominating factors which affect the food yield in Anhui province,it contains the grain acreage,effective irrigation area,the affected area,applying fertilizer content,plastic usage,applying pesticide content,the numbers of labours engaged in agricultural,forestry and fishery,total power of agricultural machinery,the numbers of existing reservoirs and AMPI.secondly,the above 11 indicators are taken as input variables,the Anhui province grain production is considered as the output variable,paper makes use of the grain yield data for 1990-2014 in Anhui province to builds the BP Neural Network,SVR,RF regression model respectively,and these models can fit the data of grain production in Anhui province.in addition,these models are evaluated and tested with 2015-2017 related data.finally,the results show that there are some defects in the imitative effect and prediction ability of each model.so,in this paper,by combining the above three models in the method of reciprocal variance effectively,combined prediction model is presented,and results show that the combined prediction model has a good fitting effect and stability on the training set.moreover,to some extent,the model adds the characteristics of global optimization,it makes the prediction results more reliable.meanwhile,the grain output of Anhui province from 2018 to 2020 is predicted to be 3459.83 million tons,3478.57 million tons and 3505.57 million tons respectively.