Codestin Search App

add the following likelihoods: ‘binomial_logit’, ‘binomial_probit’, ‘beta_binomial’, ‘negative_binomial_1’, ‘lognormal’
add cov_function = ‘linear’ (allows for doing Bayes linear regression with empirical Bayes estimation of the prior variance)
make binary classification numerically more stable
make 'binary' likelihood default to 'bernoulli_logit'
no internal normalization of ‘weights’

add likelihood = ‘beta’ (for beta regression)
faster computations for crossed grouped random effects by enabling matrix_inversion_method = “iterative” ’
add Vecchia-inducing points full-scale (VIF) approximation (gp_approx = "full_scale_vecchia")
add metric 'crps_gaussian'
add option to not estimate / hold fix some covariance parameter via the argument ‘estimate_cov_par_index’ in the ‘params’ parameter (of the ‘fit’ or ‘set_optim_params’ functions)
add ‘space_time_gneiting’ covariance function
add objective = ‘mean_scale_regression’ (for plain boosting without a GPModel)
make iterative methods as default wherever possible

faster prediction with single-level grouped random effects
shape parameter of Matern covariance can be estimated (cov_function = ‘matern_estimate_shape’, ‘matern_ard_estimate_shape’)
[R-package] make hyperparameter tuning functions more stable
use median distance instead of mean distance to initialize range parameters of GPs
change default initial values for range parameters of binary and categorical features for ARD covariance functions
change response variable prediction for 't' likelihood to avoid crashes in parameter tuning: simply add the squared 'scale' parameter to the variances of the latent predictions
make predictions with Vecchia approximations numerically more stable for non-Gaussian likelihoods

reduce memory footprint of ‘fitc’ and ‘full_scale_tapering’ GP approximations
add ‘t’ and ‘t_fix_df’ likelihoods
add ‘num_parallel_threads’ parameter so that users can control the number of parallel CPU threads
[R-package] make 'gpb.grid.search.tune.parameters' and 'tune.pars.bayesian.optimization' more robust
[python-package] make 'tune_pars_TPE_algorithm_optuna' more robust
[R-package][python-package] fix bug in handling of 'line_search_step_length' parameter in hyper parameter tuning functions
add option gp_approx = 'vecchia_latent'

support Matern covariance functions with general shape parameters (‘cov_fct_shape’), not just 0.5, 1.5, 2.5
reduce memory footprint of ‘vecchia’ approximation
"make 'fitc' and 'full_scale_taperig' approximations numerically more stable
[python-package] add function ‘tune_pars_TPE_algorithm_optuna’ for choosing tuning parameters with the Tree-structured Parzen estimator algorithm
[R-package] add function ‘tune.pars.bayesian.optimization’ for choosing tuning parameters using Bayesian optimization and the ‘mlrMBO’ R package
change default initial values for marginal variances when having multiple random effects
[R-package] fix bug in ‘gpb.grid.search.tune.parameters’ for ‘metric’ parameter for metrics for which higher is better (auc, average_precision)
other bug fixes

Support space-time (‘matern_space_time’) and anisotropic ARD (‘matern_ard’, ‘gaussian_ard’) covariance functions
support ‘negative_binomial’ likelihood
support FITC aka modified predictive process approximation (‘fitc’) and full scale approximation with tapering (‘full_scale_tapering’) with ‘cholesky’ decomposition and ‘iterative’ methods
add optimizer_cov option 'lbfgs', and make this the default for (generalized) linear effects models
faster prediction for multiple grouped random effects and non-Gaussian likelihoods
allow for duplicate locations / coordinates for Vecchia approximation for non-Gaussian likelihoods
support vecchia approximation for space-time and ARD covariance functions with correlation-based neighbor selection
support offset in GLMMs
add safeguard against too large step sizes for linear regression coefficients
change default initial values for (i) (marginal) variance and error variance to var(y)/2 for Gaussian likelihoods and (ii) range parameters such that the effective range is half the average distance
add backtracking line search for mode finding in Laplace approximation
add option ‘reuse_learning_rates_gp_model’ for GPBoost algorithm -> faster learning
add option ‘line_search_step_length’ for GPBoost algorithm. This corresponds to the optimal choice of boosting learning rate as in e.g. Friedman (2001)
support optimzer_coef = ‘wls’ when optimizer_cov = ‘lbfgs’ for Gaussian likelihood, make this the default

support iterative methods for Vecchia-Laplace approximation (non-Gaussian data and gp_approx=”vecchia”)
faster model construction and prediction for compactly supported covariance functions
add metric 'test_neg_log_likelihood'
change handling of 'objective' parameter for GPBoost algorithm: only ‘likelihood’ in ‘GPModel()’ needs to be set
change API for parameters ‘vecchia_pred_type’ and ‘num_neighbors_pred’

faster gradient calculation for
1. Multiple / multilevel grouped random effects for non-Gaussian likelihoods
2. GPs with Vecchia approximation for non-Gaussian likelihoods
3. GPs with compactly supported covariance functions / tapering
enable estimation of shape parameter in gamma likelihood
predict_training_data_random_effects: enable for Vecchia approximation and enable calculation of variances
change API for Vecchia approximation and tapering
correction in nearest neighbor search for Vecchia approximation
show GPModel parameters on original and not transformed scale when trace = true
change initial intercept for bernoulli_probit, gamma, and poisson likelihood
change default value for ‘delta_rel_conv’ to 1e-8 for nelder_mead
avoid unrealistically large learning rates for gradient descent

cap too large gradient descent steps on log-scale for covariance parameters, GLMMs: reset small learning rates for covariance parameters and regression parameters if the other parameters change
add gaussian_neg_log_likelihood as validation metric
add function ‘get_nested_categories‘ for nested grouped random effects
prediction: remove nugget variance from predictive (co)variances when predict_response = false for Gaussian likelihoods
set default value for predict_response to true in prediction function of GPModel
NA’s and Inf’s are not allowed in label
correct prediction if Vecchia approximation for non-Gaussian likelihoods

Reduce memory usage for Vecchia approximation
[R-package] add function for creating interaction partial dependence plots
Add function ‘predict_training_data_random_effects’ for predicting (=‘estimating’) training data random effects
[R-package][python-package] predict function: rename ‘raw_score’ argument to ‘pred_latent’ and unify handling of Gaussian and non-Gaussian data
(G)LMMs: better initialization of intercept, change internal scaling of covariates, change default value of ‘lr_coef’ to 0.1
Add ‘adam’ as optimizer option
allow for grouped random coefficients without random intercept effects
[R-package][python-package] nicer summary function

Releases: fabsig/GPBoost