Package index • smoothEMr

All functions
`PCA_ordering()`	PCA ordering (PC1)
`as_csmooth_em()`	Construct a csmooth_em object
`as_smooth_em()`	Construct a smooth_em object from an EM_algorithm fit
`backward_two_ordering_partition_csmooth()`	Backward greedy feature partition into two orderings (csmoothEM, warm-start only)
`compare_positions_by_history()`	Compare posterior position summaries across two histories
`do_csmoothEM()`	Run csmoothEM iterations
`do_csmoothEM_ml_collapsed()`	Run collapsed-ML csmoothEM iterations (homoskedastic only)
`do_smoothEM()`	Run SmoothEM for a given number of iterations on a smooth_em object
`fiedler_ordering()`	Fiedler ordering from a kNN graph
`forward_two_ordering_partition_csmooth()`	Forward greedy feature partition into two orderings (csmoothEM version)
`generalized_logdet()`	Generalized log-determinant (sum of log positive eigenvalues)
`get_history_block()`	Extract a (u, gamma) block for a given progressive history index
`greedy_backward_filter_csmooth()`	Greedy backward filtering of features for csmoothEM ordering inference
`hello()`	Hello, World!
`init_cov_cache_fast()`	Cache covariance inverses and log-determinants
`initialize_csmoothEM()`	Initialize csmoothEM (coordinate-specific SmoothEM)
`initialize_ordering()`	Initialize GMM parameters using an ordering method
`initialize_ordering_csmooth()`	Initialize ordering for csmoothEM (diagonal-variance version)
`initialize_smoothEM()`	Initialize SmoothEM (single- or multi-scale)
`isomap_ordering()`	Landmark Isomap ordering (no vegan; avoids O(n^2) dist matrix)
`krige_mu_list_to_full_grid()`	Krige SmoothEM mean functions from an observed grid to the final grid
`kriging_from_precision()`	Krige/interpolate from observed nodes using a Gaussian precision matrix
`lambda_scale_for_spacing()`	Scale random-walk penalty strength to account for grid spacing
`logsumexp()`	Numerically stable log-sum-exp
`make_VAR1_precision()`	Prior precision for VAR(1) on K time points (d-dimensional)
`make_default_init()`	Default random initialization for a Gaussian mixture model
`make_hierarchical_levels()`	Construct a hierarchy of nested grids inside a final grid
`make_init()`	Initialization from an ordering vector
`make_init_csmooth()`	Initialization from an ordering vector for csmoothEM
`make_lattice_rwq_precision()`	Prior precision for q-th order random walk on a K x K lattice (d-dimensional)
`make_lattice_rwq_precision_sparse()`	Sparse prior precision for q-th order random walk on a K x K lattice (d-dimensional)
`make_random_walk_precision()`	Prior precision for q-th order random walk on K time points (d-dimensional)
`make_random_walk_precision_sparse()`	Sparse prior precision for q-th order random walk on K time points (d-dimensional)
`match_locations_to_grid()`	Match locations on an old grid to indices on a new grid
`optimize_initial()`	Choose the best SmoothEM initialization by ELBO
`optimize_initial_csmoothEM()`	Optimize csmoothEM initialization by comparing multiple methods after a warm start
`parallel_initial()`	Run multiple SmoothEM initializations in parallel
`parallel_initial_csmoothEM()`	Run multiple csmoothEM initializations in parallel and summarize results
`partition_features_twofits()`	Partition features by comparing per-coordinate collapsed scores from two fits
`pcurve_ordering()`	Principal curve ordering
`plot(<csmooth_em>)`	Plot a csmooth_em object
`plot(<smooth_em>)`	Plot a smooth_em object
`plot_EM_embedding()`	Plot data embedding with SmoothEM component means
`plot_EM_embedding2D()`	Plot a 2D embedding with Smooth-EM component means overlaid
`plot_coordinate_change()`	Plot change in a single coordinate between two histories
`plot_mu_history()`	Plot kriged mean curves from mu_full_history
`plot_order_EM_overlay2D()`	Plot a 2D embedding colored by an ordering vector with EM means overlaid
`print(<csmooth_em>)`	Print csmooth_em object
`print(<smooth_em>)`	Print method for smooth_em objects
`print(<summary.csmooth_em>)`	Print summary of csmooth_em
`print(<summary.smooth_em>)`	Print method for summary.smooth_em objects
`progressive_smoothEM()`	Progressive-resolution initialization for SmoothEM via kriging on a final grid
`score_feature_given_Gamma()`	Score a single feature given responsibilities (Gamma)
`score_features_onefit()`	Score features under a fitted csmooth_em model via per-coordinate collapsed contributions
`simulate_spiral2d()`	Simulate a 2D Archimedean spiral (helix-like) with noise
`simulate_swiss_roll_1d_2d()`	Simulate a 2D "swiss roll" spiral with a 1D latent parameter
`simulate_two_order_gp_dataset()`	Simulate a two-ordering GP dataset (Matern) for feature partitioning
`subset_csmooth_em_fit()`	Subset a csmooth_em object by features (columns)
`summary(<csmooth_em>)`	Summary for csmooth_em object
`summary(<smooth_em>)`	Summary method for smooth_em objects
`tSNE_ordering()`	t-SNE ordering

Reference

All functions