fdaPDE 1.1-17

New features

  1. Wald inference on the nonparametric term f for space-time models

  2. Spatio-Temporal Density Estimation

fdaPDE 1.1-9

New features

  1. Inference in space and space-time model (Gaussian family, no forcing terms): In space model inference is available for both parametric and nonparametric components; In space-time model inference only on the parametric component has been implemented.

fdaPDE 1.1-4

New features

  1. fdaPDE over Linear Networks domains
  2. smooth regression for non-gaussian space-time data (GLM model)

fdaPDE 1.1-3

New features

Optimization in spatio-temporal setting: introduced Newton method and refactored grid code.

fdaPDE 1.1-2

New features

Iterative method for spatio-temporal regression with PDE regularization (in this framework also two methods for the computation of GCV Exact and Stochastic)

fdaPDE 1.1-1

New features

  1. Optimization methods (Newton’s methods) to find best smoothing parameter through GCV minimization
  2. smooth regression for non-gaussian data (GLM model)

fdaPDE 1.1-0

New features

  1. smooth regression for space-time data
  2. density estimation
  3. faster search algorithm

fdaPDE 1.0-8

Bug fixes

Compilation errors with clang fixed.

fdaPDE 1.0-7

Bug fixes

Compilation error in macOS fixed.

fdaPDE 1.0-6

Bug fixes

Compilation error in solaris fixed.

fdaPDE 1.0-5

Bug fixes

gcc-ASAN problem with RTriangle fixed.

fdaPDE 1.0-1

Bug fixes

Compilation errors with clang fixed.

fdaPDE 1.0

New features

  1. smooth regression for manifold domains
  2. smooth regression with areal data
  3. functional smooth PCA (SF-PCA algorithm) : function FEM.FPCA
  4. Stochastic GCV computation has been added: parameter ‘GCVmethod’ can be used both regression and FPCA, can be either ‘Stochastic’ or ‘Exact’.
  5. Kfold cross-validation is available in FPCA algorithm.

Deprecated functions and name changes

  1. smooth.FEM.basis, smooth.FEM.PDE.basis and smooth.FEM.PDE.sv.basis are deprecated, use smooth.FEM instead.
  2. the R-only functions (whose names began with R_) have been deprecated, and will be removed soon.
  3. The usage of ‘MESH’ in classes and function names has been deprecated, now use ‘mesh’.
  4. Old datasets have been removed. New datasets are provided.
  5. The parameter CPP_CODE has been removed.

Bug fixes

A bug in R/C++ indexes conversion in the space-varying regression has been fixed.