Research On The Application Of Cotton Yield Prediction Model Based On Grey Neural Network

Cotton is an important crop for the country’s economic development and the main raw material for the development of the cotton textile industry.The cotton industry is an important pillar industry of the country’s national economy.Xinjiang is the main cotton production base in the country.In 2021,Xinjiang’s cotton planting area will be 37.5926 million mu,accounting for 82.8% of the national cotton planting area,and its cotton output will be 5.129 million tons,accounting for 89.5% of the national cotton output.Xinjiang cotton occupies an important position in the development of the national cotton textile industry.Cotton is affected by many factors in the production process.In addition to varieties,factors such as natural environment,fertilizer application,soil fertility,agricultural electricity consumption,total power of agricultural machinery and planting area will have a very important impact on cotton production.Therefore,it is of great practical significance to analyze the key factors affecting cotton yield,establish a cotton yield prediction model,and scientifically and accurately predict Xinjiang cotton yield for guiding agricultural production and promoting my country’s economic development.This paper takes cotton as the research object and selects the annual cotton yield and related influencing factors data in Xinjiang Uygur Autonomous Region from 1990 to 2019,and uses the grey forecast model and neural network model to study the cotton yield forecast in Xinjiang.First,12 index factors affecting cotton yield were determined from the four perspectives of planting area and labor force,agricultural production input,climate and natural disasters,and technology and government support,and the relationship between the 12 influencing factors and cotton yield was calculated by grey relational analysis method.Correlation degree,remove the two influencing factors with low correlation degree,and select the remaining ten influencing factors to establish an index system suitable for predicting cotton yield.Secondly,according to the complex characteristics of cotton production,combined with the pros and cons of forecasting algorithms,the grey forecasting model,the BP neural network model and the two combined models of the two(series grey neural network and parallel grey neural network)were selected to predict the annual output of cotton in Xinjiang.Finally,the average relative error(MRE),root mean square error(RMSE)and mean absolute error(MAE)are used to evaluate the performance of the above four models.The model is used to forecast the annual cotton output in Xinjiang from 2020 to 2022,which are 5.3731 million tons,5.3203 million tons and 5.3829 million tons respectively.