基于优化极限学习机模型的京津冀地区气象干旱预报研究

Research on meteorological drought forecast in Beijing-Tianjin-Hebei region based on optimized extreme learning machine model

基于京津冀地区气象干旱较严重的现状,为找出适用于京津冀地区干旱预报的标准模型,以相对湿润指数(MI)和极限学习机模型(ELM)为基础,基于麻雀搜索算法(SSA)、粒子群算法(PSO)、遗传算法(GA)3种优化算法,构建了SSA-ELM、PSO-ELM、GA-ELM共3种优化模型,并将计算结果与ELM模型、广义回归神经网络模型(GRNN)和BP神经网络模型作比较,结果表明:京津冀地区的气象干旱普遍较严重,尤其在春季和冬季,全区基本均以特旱为主;SSA-ELM模型在干旱预报中表现出了较高的精度,该模型的误差指标最低,同时一致性指标最高,且综合性绩效指数(GPI)为1.36,在所有模型中精度排名第1,因此,SSA-ELM模型可作为京津冀地区干旱预报的推荐模型使用。

Based on the serious drought situation in the Beijing-Tianjin-Hebei region, to find a standard model for drought forecasting in the region, we adopted the principle of machine learning model, based on the relative humidity index(MI) and extreme learning machine model(ELM), and three optimization algorithms including sparrow algorithm(SSA), particle swarm algorithm(PSO) and genetic algorithm(GA). Three optimization models(SSA-ELM, PSO-ELM and GA-ELM) were constructed. The calculation results were compared with ELM model, generalized regression neural network model(GRNN) and BP neural network model. The results showed that: the degree of drought in the BeijingTianjin-Hebei region was generally serious, especially in spring and winter, and the whole region was basically dominated by extreme drought. The GPI of the SSA-ELM was 1.36, ranking 1st among all the models. The SSA-ELM model can be used as a recommended model for drought forecasting in the Beijing-Tianjin-Hebei region.