Research On Early Prediction Of Apple Yield In Yuncheng City Based On Meteorological Data

Apple is an important economic crop and is widely planted in my country.It is loved by people because of its high nutritional value and sweet taste.Apple planting is a pillar industry in Yuncheng City.Apple production is of great significance to promoting the economic development of Yuncheng City.The realization of accurate prediction of apple production is of great significance to preventing market fluctuations caused by changes in apple production.In the apple production process,meteorological factors are closely related to apple production.Therefore,this paper uses HP filtering method to divide the annual apple output of Yuncheng into trend apple annual output and meteorological apple annual output.On this basis,it deeply studies the influence of various meteorological factors on apple meteorological output.In addition,the traditional apple production forecast is to take a certain method to estimate the output after the apple is mature and before the harvest,which has a certain lag in guiding the apple market circulation.This paper analyzes the meteorological data of Yuncheng City from 2005 to 2018 and the annual output data of apples,and proposes an early prediction method of Yuncheng apple output based on meteorological data.Using the meteorological data of the young apple fruit period,the BP neural network model is established to realize the comparison.Early forecast of apple production.The main work of the paper is as follows:1.The study found that the annual output of apples is mainly composed of trend output(the long-period output component reflecting the level of productivity development in the historical period)and meteorological output(the fluctuating output component affected by the short-period change factors dominated by climatic factors).The HP filter method divides the annual output of apples in Yuncheng City into annual output of trending apples and annual output of meteorological apples.On this basis,the influence of various meteorological factors on the meteorological output of apples is deeply studied.2.Multiple linear regression model is established by using the meteorological data of this period and apple meteorological yield data of apple phenological period: germination period,flowering period,young fruit period,swelling fruit period and maturity period to study the influence of each phenological period on apple meteorological yield,and to find the most critical period affecting apple yield,respectively.According to the results of multiple linear regression models established for each phenological stage of apple,the goodness of fit(2R)of multiple regression models for each phenological stage is as follows:germination stage 0.814,flowering stage 0.871,young fruit stage 0.884,rising fruit stage 0.730,and maturity stage 0.826.By comparison,it can be seen that the model of young fruit stage has the highest goodness of fit,and the correlation between meteorological factors and meteorological yield is the closest during this period,indicating that young fruit stage is the most critical period affecting apple yield.3.The 11 meteorological factors that had the strongest effect on the apple meteorological yield at the young fruit stage: Maximum air pressure,minimum air pressure,maximum temperature,minimum temperature,average humidity,minimum humidity,rainfall,average wind speed,maximum wind speed,sunshine duration and average ground temperature are input to the model,and annual apple production data are taken as the model output.The BP neural network apple production early prediction model is established and verified.According to the training of the early prediction model of apple yield of BP neural network,the average relative error of the prediction results is 7.08%.After the relevant data of 2019 are substituted into the model for verification,the model accuracy is 89.6%,indicating that the model is suitable for the early production prediction of Apple.However,the relative error between the actual yield and the predicted yield in some years of the model is large,which still needs to be improved.4.The input parameters of the model are optimized,and the meteorological factors with high correlation with apple yield are found by using Grey correlation analysis,which are taken as the input of the model.Genetic algorithm is used to optimize BP neural network,and the early prediction model of BP neural network is established after optimization.Finally,the model results before and after optimization are compared and analyzed.The Grey correlation analysis results show that the maximum temperature,minimum temperature,average humidity,rainfall,maximum wind speed,sunshine duration,average ground temperature and other meteorological factors are highly correlated with the apple yield,and have a strong effect on the apple yield.The average relative error of the prediction results of the optimized BP neural network model is 1.23%,and there is little difference between the actual yield of samples and the relative error of predicted yield in each year.After the relevant data of 2019 are substituted into the model for verification,the accuracy of the model is 97.3%,indicating that this model is more feasible and effective for the early prediction of Apple production compared with the BP neural network model.In conclusion,apple young fruit stage is the key period affecting apple yield,fruit farmers should pay attention to young fruit stage management.The optimized BP neural network model proposed in this paper can predict the apple yield more accurately,which can provide theoretical support for the early prediction of apple yield.