智慧农业导刊2024，Vol.4Issue(11)：11-15,5.DOI:10.20028/j.zhnydk.2024.11.003
PSO-SVR模型在吉林省干旱指数预测中的应用
摘要
Abstract
With the escalating severity of climate change,drought has emerged as a significant threat to agricultural production in China,severely impeding the development of the agricultural sector.Therefore,mastering the technology for scientifically predicting drought indices is crucial.This technology can provide preventive recommendations for future drought situations,prevent the further expansion of drought,and consequently ensure food security.Based on meteorological factors such as air pressure,temperature,precipitation,relative humidity,and the Standardized Precipitation Evapotranspiration Index(SPEI)in 40 regions of Jilin Province,the future drought indices were forecasted.The four error indicators of the BP model,RF model,SVR model,and the optimized PSO-SVR model were compared,which reveals that the PSO-SVR model demonstrated superior performance.With an R2 reaching 0.964 and MSE at 0.021,it outperformed the other three models,exhibiting a more significant fitting effect.The results indicate that the PSO-SVR model exhibits high feasibility and accuracy in predicting the SPEI index in Jilin Province,providing robust theoretical support for drought prevention and mitigation research in the region.关键词
SPEI/SVR/预测模型/干旱指数/回归分析Key words
SPEI/SVR/prediction model/drought index/regression analysis分类
农业科技引用本文复制引用
徐子曦,钟闻宇,唐友..PSO-SVR模型在吉林省干旱指数预测中的应用[J].智慧农业导刊,2024,4(11):11-15,5.基金项目
吉林省科技发展计划项目(YDZJ202201ZYTS692) （YDZJ202201ZYTS692）
吉林农业科技学院横向课题(横20230052) （横20230052）
