Monitoring And Forecasting Model Establishment Of Agriphila Aeneociliella,A New Insect Pest Of Wheat

China is a large agricultural country.The control of agricultural pests is particularly important for ensuring national food safety and effective supply of main agricultural products.In recent years,big data processing technology has gradually applied to agricultural field to promote agricultural transformation and upgrading.Based on big data,the occurrence amount of insect pests in the past years is analyzed,and the main factors that associated with the outbreak of the pests can be revealed,so as to make scientific prevention and control decisions and reduce the grain loss.According to the meteorological data of the Agriphila aeneociliella occurrence period in eastern Shandong,the A.aeneociliella monitoring and forecasting model is established based on R program,and the key meteorological factors for its outbreak is revealed.The damage extent is predicted.The results are as follows:1.Using random forest algorithm based on R program,analyzing the relationship between the occurrence degree and the meteorological factors from 2010 to 2017 in eastern Shandong,with 17 meteorological factors as independent variables,the occurrence degree of A.aeneociliella as the dependent variable,the monitoring and forecasting model was established and the main meteorological factors affecting the occurrence of A.aeneociliella were figured out.According to the model analysis,the meteorological factors including‘ Average surface temperature',‘average vapor pressure',and the ‘daily minimum temperature',played a major role in affecting the occurrence degree of A.aeneociliella in eastern Shandong province,the mean decrease Gini were 14.56,13.35 and 11.74,respectively.The accuracy of the models was up to 81.14% or 81% depending on out of bag estimation or text data.This model could provide the technical support for the occurrence monitoring and forecasting of A.aeneociliella.2.Using R program analysis platform,multivariate statistical analysis was conducted based on the occurrence of A.aeneociliella and meteorological variables in eastern Shandong.A principal component regression analysis model for the pest is constructed.The principal components analysis is used to reduce the dimension and finally get 5 principal components.The cumulative contribution rate is 86.46%.Then the 5 principal components are as independent variables for multiple regression.The multiple R-Squared is 0.8114.According to the multiple collinearity test and residual plot test,this regression model is reliable to forecast the occurrence degree of A.aeneociliella.