关键词: 干旱预测, GA-ELM, SPEI, 云贵高原,

Abstract: In order to accurately predict drought and improve drought control and relief ability, the genetic algorithm-optimized extreme learning machine model(GA-ELM) was constructed to predict drought. In this paper, key drought-producing factors were used as input variables to realize the medium and long-term drought prediction of the Yunnan-Guizhou Plateau by GA-ELM. Then, the results by GA-ELM was compared with that of adaptive neural fuzzy inference system(ANFIS) model and extreme learning machine(ELM) model. The results showed that the GA-ELM could be applied to drought prediction in Yunnan-Guizhou Plateau. Also, the results obtained by GA-ELM were more accurate than that of ELM at different time scales. The prediction accuracy of GA-ELM was higher than that of ANFIS in terms of drought intensity and drought duration.

Key words: drought prediction, GA-ELM, SPEI, Yunnan-Guizhou Plateau