The infestation of Drosophila melanogaster seriously affects blueberry output and has become one of the main reasons to curb the development of blueberry industry in Guizhou Province.Rapid and accurate prediction of D.melanogaster occurrence is beneficial for timely taking prevention and control measures.How-ever,there are currently few researches on occurrence prediction of D.melanogaster in blueberry orchards.Therefore,a model for predicting the occurrence of D.melanogaster in blueberry orchards was proposed in this study.Firstly,the Pearson correlation coefficient was used to analyze the correlations between meteorological factors such as temperature,humidity and wind speed and occurrence of D.melanogaster.Then,the Random Forest algorithm was used to select important climate features that affect the occurrence of D.melanogaster.Fi-nally,a pest prediction model combined with Random Forest and Long Short-Term Memory Network was pro-posed.Comparing the predictive performance of this model with other traditional models,the results showed that it performed well in D.melanogaster occurrence prediction with the root mean square error as 2.120 3,the average absolute error as 1.865 9,and the coefficient of determination as 0.979 5.The results of this study could provide technical supports for monitoring the occurrence of D.melanogaster and adopting corresponding prevention and control strategies in time.
关键词
黑腹果蝇/蓝莓/虫害发生预测/随机森林/长短期记忆网络/Pearson相关系数/气候特征
Key words
Drosophila melanogaster/Blueberry/Prediction of pest occurrence/Random Forest/Long Short-Term Memory Network/Pearson correlation coefficient/Climate features
维普
万方数据