不明觉厉的研究

  • GNN

    GNN是一种深度学习模型，通过聚合节点的邻居嵌入来生成节点嵌入

    • Spatio-temporal graph neural networks (STGNN)

      GNN的一个类别，常用于交通流量预测、流行病预测和睡眠阶段分类

      • 方法：基于CNN和基于RNN
      • 图卷积

    • Graph Diffusion Convolution

  • Diffusion Convolution

  • Denoising Diffusion Model

  • Time Series Forecasting

Time Series Forecasting

Time series forecasting aims to predict future time series values based on historical observations.

Method

  • RNNs, CNNs, attention-based

    Overlook and implicitly model the rich underlying dynamic spatial correlations between time series.

  • GNNs

    have shown great potential in explicitly and effectively modeling spatial and temporal dependencies in multivariate time series data

金融领域

  • 金融时序

    Financial time series usually come with a high degree of variability and extensive noise to disrupt the forecasting performance.

  • 金融预测方法

    • 对抗生成网络
    • GNN
      • GCN+GRU（图卷积网络+门控循环单元）
      • STGNN
      • HGN（超图网络）
      • R-GCN（关系图卷积网络）
      • Multi-graph Convolutional Network
    • LSTM-ReGAT（长短期记忆网络-关系图注意力网络）