Module function_on_scalar_2d 

2D Function-on-Scalar Regression (FOSR).

Extends the 1D function-on-scalar model to surface-valued functional responses observed on a regular 2D grid:

Y_i(s,t) = beta_0(s,t) + sum_j z_{ij} beta_j(s,t) + epsilon_i(s,t)

The estimation uses a two-step approach:

1. Pointwise OLS at each grid point to obtain raw coefficient surfaces.
2. Tensor-product roughness penalty smoothing of each coefficient surface.

Methods

• fosr_2d: Penalized 2D FOSR with anisotropic smoothing
• predict_fosr_2d: Predict new surfaces from a fitted model

Structs

FosrResult2d

Result of 2D function-on-scalar regression.

Grid2d

2D grid description for surface-valued functional data.

Functions

fosr_2d

2D Function-on-Scalar Regression with tensor-product penalty.

predict_fosr_2d

Predict functional surfaces for new observations.