基于组合赋权模型的重庆降雨预测模型构建

Construction of Chongqing Rainfall Prediction Model Based on Combined Weighting Model

为找出适用于区域不同尺度降雨的估算模型,本研究以重庆为研究区域,以极限学习机模型(ELM)为基础,采用鸽群算法(PIO)和蝙蝠算法(BA)对ELM模型进行优化,与XGBoost模型进行组合赋权,得出了XG-PIO-ELM和XG-BA-ELM共2种组合赋权模型,得出了适用于重庆地区的降雨估算模型,结果表明:XG-PIO-ELM在模拟重庆降雨日值和月值时均表现出了较高的精度,可推荐用于估算重庆降雨。

In order to find an estimation model for rainfall at different scales in the region,in this research,we took Chongqing as the research area,based on the Extreme Learning Machine Model(ELM).We used Pigeon Swarm Algorithm(PIO)and Bat Algorithm(BA)to optimize the ELM model,and performed the combined weighting with XGBoost model.Two combined weighting models,XG-PIO-ELM and XG-BA-ELM,were obtained,resulting in a rainfall estimation model suitable for Chongqing region.The results showed that:XG-PIO-ELM showed the highest accuracy in simulating daily and monthly rainfall values in Chongqing,and can be recommended for estimating rainfall in Chongqing.