【摘要】 近年来各种降雨导致的自然灾害频繁发生，给人们的日常生活带来较大影响，及时准确的短期降雨预测可以提醒人们做好预防措施，然而影响短期降雨的天气因素多且变化快，难以对其进行准确预测。对此提出一种基于BiGRUCapsNet与Transformer的双分支短期降雨预测模型，将预处理好的数据分别输入BiGRU-CapsNet与Transformer进行特征提取，然后将提取的特征融合后输入到全连接层进行短期降雨预测。实验结果表明，所提模型在准确率、精准率、F1分数等评价指标均取得较好的结果，能够对短期降雨进行较准确预测。更多还原

【Abstract】 In recent years,various natural disasters caused by rainfall occur frequently,which have a great impact on People’s daily life. Timely and accurate short-term rainfall prediction can remind people to take preventive measures. However,the weather factors affecting short-term rainfall are many and change quickly,so it is difficult to accurately predict short-term rainfall. In this paper,a dual-branch short-term rainfall prediction model based on BiGRU-CapsNet and Transformer is proposed. The preprocessed data are respectively input into BiGRU-CapsNet and Transformer for feature extraction,and then the extracted features are fused into the fully connected layer for short-term rainfall prediction. The experimental results show that the proposed model achieves good results in the evaluation indexes such as accuracy,precision and F1 score,and can accurately predict short-term rainfall.更多还原