mixgb()xgb.train():num_class now passed throughparams list,need to setnum_class=NULL for non-multiclassimputation.mixgb_null() function for better readabilityand maintainability - using helper functions inimpute_each.R. Other relevant functions will be updatedlater.-framework Accelerate flag forLinux builds inMakevars file ### related to XGBoostParameter Changesxgb.train():num_class now passed throughparamslist.reshape = TRUE argument inpredict() function as it is deprecated in XGBoost >=2.0.0.xgb.train():watchlist is changed toevalsobjective andeval_metric are now passedthroughparams list.xgb.cv():sparse.model.matrix intoxgb.DMatrix format. Note, this is a quick fix for minimalsafe change, and we plan to further optimise data handling in futurereleases.best.nrounds is now obtained viacv.train$early_stop$best_iteration instead ofcv.train$best_iteration.cbind2(Mis.m, Obs.m) that occurred whenthe imputed dataset had only a single incomplete variable.Mis.m (a matrix ofall other incomplete variables except the currently imputed one) was a0x0 matrix, which caused error during binding.vismi packagedrop.unused.levels = FALSE infac2sparse() to prevent dropping unused levels in factor orordinal factor.save_yhatobs() for Type 1 pmm.mixgb() for large datasets:mixgb(). Users canstill use bootstrap in the archived functionmixgb0().PMM is now set toNULL by default.xgb.save() andxgb.load() from XGBoost.device.gpu_id andpredictor.tree_method = "hist" by default, aligning withXGBoost 2.0.0.save.models.folder inmixgb().xgb.save(), a methodrecommended by XGBoost for future compatibility.save.models.folder is specified, the return objectofmixgb() includes the current imputed datasets,directories for imputation models, and relevant parameters. This objectcan save usingsaveRDS() as it doesn’t directly contain themodels. Users can later load this object into R and employimpute_new(object, newdata, ...) for new dataimputation.mixgb(data,...) to support datasets withdiverse data types:Note: Users must manually convert character variables to factors.
default_params(), an auxiliary function formixgb(), to validate the list of XGBoost hyperparameterssupplied by the user. It simplifies hyperparameter modifications withoutrequiring explicit specification of all default values.plot_hist() andplot_bar() toalign with changes in ggplot2 3.4.0:..density.. withafter_stat(density) inplot_hist()...prop.. withafter_stat(prop) inplot_bar().nthread = 2 to comply with CRANpolicies.subsample = 0.7, becomes the default method due toidentified biases with bootstrapping in certain scenarios.mixgb():subsample = 0.7.bootstrap = FALSE.createNA() function.mixgb():ordinalAsInteger: Changes fromTRUE toFALSE.max_depth: Changes from 6 to 3.nrounds: Changes from 50 to 100.bootstrap: Sets toTRUE by default.mixgb asxgboost requires OpenMP formulti-core operations. For details, please refer toOpenMP for Mac.