The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

missMethods 0.4.0

New functions

• count_NA() counts the number of missing values in a vector, data frame or matrix.

Improvements

• apply_imputation() and friends now have an option to convert tibbles instead of throwing an error.

Fixes

• delete_MAR_1_to_x() and delete_MNAR_1_to_x() now handle unordered factors as documented (thanks to Steve Roehrig for reporting).
• delete_MAR_1_to_x() and delete_MNAR_1_to_x() now display the correct adjusted x value, if it is too high or too low (thanks to Steve Roehrig for reporting).
• apply_imputation() type rowwise now works for data frames (thanks to @khughitt for fixing).

Internal

• delete_values() now only takes mech_type and derives mechanism.

missMethods 0.3.0

Update of delete_ functions

User-visible changes

• All delete_ functions have the argument n_mis_stochastic now. For some functions, this is only a renaming of the old stochastic argument (e.g. delete_MCAR()), for others this is completely new. The new name emphasis that this argument controls if the number of missing values is stochastic or deterministic.
• delete_MAR_1_to_x() and delete_MNAR_1_to_x() get a new argument x_stochastic along the line of n_mis_stochastic.
• Add the (package-wide) option missMethods.warn.too.high.p to control the displaying of warnings for too high values of p (the probability for a value to be missing).

Internal

• delete_values() and get_NA_indices() centralize many steps of the old (not exported) delete_ functions.
• All delete_MAR_ and delete_MNAR_ functions and delete_MCAR() call the new delete_values() function now.
• Most of the delete_ functions use the new get_NA_indices() to determine the missing values.

Miscellaneous

• Evaluation functions can now compare a data frame with a matrix (thanks to Marie Feldhoff for the suggestion).
• impute_EM() now returns the number of performed EM iterations as attribute.
• delete_rank() now hands the argument ties.method over to rank().
• Fix two tests for delete_one_group() (wrong argument FUN instead of cutoff_fun).
• Correct documentation of median.factor() (thanks to @labachevskij).
• impute_LS_adaptive() has now the default setting warn_r_max = FALSE.
• Remove LazyData from DESCRIPTION (fix for a CRAN NOTE).

missMethods 0.2.0

New functions

• impute_in_classes() allows to apply any imputation function inside imputation classes
• impute_hot_deck_in_classes() hot deck imputation inside of imputation classes (adjustment cells)
• impute_EM() imputes values using EM parameter estimates
• imputed_expected_values() imputes expected values from a multivariate normal distribution
• impute_LS_adaptive() performs LSimpute_adaptive as described by Bo et al. (2004)
• impute_LS_array() performs LSimpute_array as described by Bo et al. (2004)
• impute_LS_combined() performs LSimpute_combined as described by Bo et al. (2004)
• impute_LS_gene() performs LSimpute_gene as described by Bo et al. (2004)

Miscellaneous

• add cov_only and cor_only as parameter in evaluate_imputation_parameters()
• improve vignette “Generating missing values” (add connections to Santos et al. (2019))
• rename of cols variables: now all should be named cols_mis, cols_ctrl etc.
• rename of ds variables: now all should be named ds_imp, ds_orig etc.
• rename of pars variables: now all should be named pars_est or pars_true
• sampling of sRHD type cols_seq is now correct, if the donor is only one numeric value
• use markdown for documentation of new functions

missMethods 0.1.0

New functions

Functions for the creation of missing values:

• delete_MAR_censoring() and delete_MNAR_censoring() create missing (not) at random values using a censoring mechanism
• delete_MAR_one_group() and delete_MNAR_one_group() create missing (not) at random values by deleting values in one of two groups
• delete_MAR_rank() and delete_MNAR_rank() create missing (not) at random values using a ranking mechanism

Functions for evaluation:

• evaluate_imputation_parameters() compares estimated parameters after imputation to true parameters

New features

• delete_MAR_1_to_x() and delete_MNAR_1_to_x() can now handle (unordered) factors
• new criteria for evaluate_imputed_values() and evaluate_parameters(): six forms of NRMSE, nr_equal, nr_NA and precision
• evaluate_imputed_values(): add argument cols_which to select columns for evaluation.

Miscellaneous

• all delete_ functions now take the same first three arguments: ds, p, cols_mis
• package now on GitHub and CRAN

missMethods 0.0.1

Implemented functions

Functions for the creation of missing values:

• delete_MCAR() creates missing completely at random values in different ways
• delete_MAR_1_to_x() and delete_MNAR_1_to_x() create missing (not) at random values using a 1:x mechanism

Functions for imputation:

• impute_mean(), impute_median(), impute_mode() different forms of mean, median and mode imputation
• impute_sRHD() simple Random Hot-Deck imputation with the possibility to specify a donor limit
• apply_imputation() a function to apply aggregating functions for imputation

Functions for evaluation:

• evaluate_imputed_values() compares imputed to true values
• evaluate_parameters() compares estimated to true parameters

Miscellaneous:

• median.factor() computes medians for ordered factors

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.