Baseline methods
The baseline methods solve the temperature field reconstruction problem from four different ways. [Code]
Point-based Methods
- k-nearest nonlinear interpolation(KInterpolation)
- Global Gaussian interpolation(GInterpolation)
- Polynomial regression(PR)
- Random forest regression(RFR)
- Gaussian process regression(GPR)
- Support vector regression(SVR)
- MLP for point-based modelling(MLPP)
- Restricted boltzmann machines(RBM)
- Deep belief networks(DBNs)
Vector-based Methods
- MLP for vector-based modeling (MLPV)
- Conditional Neural Process (CNP)
- Transformer
Image-based Methods
- Fully convolutional networks (FCN)
- UNet
- SegNet
- Feature pyramid networks (FPNs)
Graph-based Methods
- Graph convolutional networks (GCNs)