基于ELM的日降雨预测模型研究

采用极限学习机(ELM)对广西降雨历史资料进行训练建模预测,并与同期的T213(中国气象局的全球中期天气数值预报产品预报值)模式预报进行对比,实验结果表明,预测均方误差及最大绝对误差均优于同期的T213模式预报,模型有一定的参考价值。

The extreme learning machine(ELM)is used to model and forecast the historical rainfall data in Guangxi,and compared with the T213 model forecast in the same period.The experimental results show that the prediction mean square error precision is better than the T213 model forecast in the same period.