目录
- 【2021 NeurIPS】Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecasting
- 【2022 ICML】FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecasting
- 【2023 ICLR】TIMESNET: TEMPORAL 2D-VARIATION MODELING FOR GENERAL TIME SERIES ANALYSIS
【2021 NeurIPS】Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecasting
关键词:时序分解模块、自相关机制
特点:注重时序数据自身的周期特征
自测:使用自己的时序数据集进行测试,效果不错
【2022 ICML】FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecasting
关键词:频域、傅里叶变换、小波变换
特点:在Autoformer主体框架的基础上,使用傅里叶变换、小波变换将时域转换为频域后进行特征提取
自测:暂未测试
【2023 ICLR】TIMESNET: TEMPORAL 2D-VARIATION MODELING FOR GENERAL TIME SERIES ANALYSIS
关键词:升维
特点:将一维时序数据转化为二维数据,借助CV领域成熟模型进行特征提取
自测:使用自己的时序数据集进行测试,效果显著