Kolmogorov-Arnold Networks with Interchangeable Edge Bases Using 'torch'


[Up] [Top]

Documentation for package ‘thangldKANr’ version 0.1.0

Help Pages

chebyshev_basis Chebyshev polynomials and their derivatives
chebyshev_basis_deriv Chebyshev polynomials and their derivatives
hybrid_basis Basis functions of a hybrid KAN edge
hybrid_basis_deriv Basis functions of a hybrid KAN edge
hybrid_kan_components Size of each basis part on data
jacobi_basis Jacobi polynomials and their derivatives
jacobi_basis_deriv Jacobi polynomials and their derivatives
kan_bases List the available edge bases
kan_layer_bases Bases used by the layers of a model
legendre_basis Legendre polynomials and their derivatives
legendre_basis_deriv Legendre polynomials and their derivatives
mexican_hat Mexican hat (Ricker) wavelet
mexican_hat_deriv Mexican hat (Ricker) wavelet
plot.thangld_kan Plot a Kolmogorov-Arnold network
predict.thangld_graph_kan Predict node targets
predict.thangld_kan Predict with a fitted Kolmogorov-Arnold network
predict.thangld_temporal_kan One-step-ahead forecast
rational Safe Pade rational function
rational_deriv Safe Pade rational function
rbf_basis Gaussian radial basis functions
rbf_basis_deriv Gaussian radial basis functions
rbf_grid Gaussian radial basis functions
register_kan_basis Register an additional edge basis
relu_basis ReLU-KAN basis functions and their derivatives
relu_basis_deriv ReLU-KAN basis functions and their derivatives
thangld_graph_kan Create a graph KAN architecture
thangld_graph_kan_adjacency Normalized adjacency matrix of an undirected graph
thangld_graph_kan_fd_gradient_r Forward pass and gradients in base R
thangld_graph_kan_fit Fit a graph KAN for node regression
thangld_graph_kan_forward_r Forward pass and gradients in base R
thangld_graph_kan_gradient_r Forward pass and gradients in base R
thangld_graph_kan_load Save and load a graph KAN
thangld_graph_kan_metrics Regression metrics
thangld_graph_kan_save Save and load a graph KAN
thangld_graph_kan_simulate Simulate a stochastic block model graph with node features
thangld_graph_kan_validate Run the numerical validation suite
thangld_kan Create a Kolmogorov-Arnold network with a chosen edge basis
thangld_kan_edge Evaluate a learned edge function
thangld_kan_fd_gradient_r Forward pass and gradients in base R (numerical oracle)
thangld_kan_fit Fit a Kolmogorov-Arnold network for regression
thangld_kan_forward_r Forward pass and gradients in base R (numerical oracle)
thangld_kan_gradient_r Forward pass and gradients in base R (numerical oracle)
thangld_kan_load Save and load a Kolmogorov-Arnold network
thangld_kan_metrics Regression metrics
thangld_kan_save Save and load a Kolmogorov-Arnold network
thangld_kan_validate Run the numerical validation suite
thangld_temporal_kan Create a lag-window Temporal KAN
thangld_temporal_kan_embed Lag-window design matrix
thangld_temporal_kan_fit Fit a Temporal KAN on a series
thangld_temporal_kan_load Save and load a Temporal KAN
thangld_temporal_kan_rolling Rolling-origin one-step evaluation
thangld_temporal_kan_save Save and load a Temporal KAN
thangld_temporal_kan_validate Run the numerical validation suite
unregister_kan_basis Register an additional edge basis