vimp - Fix documentation NOTEs

Fix CRAN check NOTEs in documentation due to new \link{} handling, and some new NOTEs due to differences between \itemize and \describe.

- R
Published by bdwilliamson 7 months ago

vimp - Add cluster bootstrap

Add a cluster bootstrap for correlated data.

- R
Published by bdwilliamson over 2 years ago

vimp - Enhanced VIM point estimation

For inference on VIM values to be valid under the zero-importance null hypothesis, we need this inference to be based on predictiveness (of the two groups of variables that we're comparing) that is estimated on independent splits of the data.

However, for point estimation, this is not required. Until now, the only option has been for the final VIM point estimate to be based on this sample-splitting. Now, you have the option for the final point estimate to be based on the entire dataset (setting final_point_estimate = "full") or for the estimate to be based on averaging the two split-specific VIMs (setting final_point_estimate = "average").

In large datasets (with many rows), there will not be much difference between these three options. However, using more data may help stabilize the final point estimate in smaller sample-size settings. Care must be taken in interpretation, however: inference is still based on sample-splitting, so an identity-link Wald confidence interval will not be centered at the final point estimate if "full" or "average" is used.

- R
Published by bdwilliamson about 3 years ago

vimp - vimp 2.3.0

Major changes

• Predictiveness measures now have their own S3 class, which makes internal code cleaner and facilitates simpler addition of new predictiveness measures.
• In this version, the default return value of extract_sampled_split_predictions is a vector, not a list. This facilitates proper use in the new version of the package.

Minor changes

• You can now specify truncate = FALSE in vimp_ci

- R
Published by bdwilliamson over 3 years ago

vimp - Specify 'method' and 'family' in outer functions

• Specify method and family for weighted EIF estimation within outer functions (vim, cv_vim, sp_vim) rather than the measure* functions. This allows compatibility for binary outcomes.
• Added a vignette for coarsened-data settings.

- R
Published by bdwilliamson over 3 years ago

vimp - Allow parallel computation of CV.SuperLearners

Allow parallel computation in CV.SuperLearner, but not in SuperLearner.

- R
Published by bdwilliamson about 4 years ago

vimp - Improved bootstrap and coarsened-data behavior

Allow further specification of the bootstrap (e.g., percentile); update documentation and internal checks for coarsened-data settings.

- R
Published by bdwilliamson over 4 years ago

vimp - Allow odd number of folds for CV-VIM with precomputed regressions

- R
Published by bdwilliamson over 4 years ago

vimp - Add data

In this release, we: - added a dataset, vrc01, containing neutralization sensitivity values to the broadly neutralizing antibody VRC01 and HIV viral sequence features - updated vignettes to use vrc01 - revised how odd numbers of cross-fitting folds are handled

- R
Published by bdwilliamson over 4 years ago

vimp - Finalize standard errors

Updated computation of standard errors. This release is harmonized with CRAN release 2.2.3.

- R
Published by bdwilliamson over 4 years ago

vimp - Harmonize with CRAN

Pass CRAN checks and harmonize with cvAUC, ROCR.

- R
Published by bdwilliamson over 4 years ago

vimp - Improving power

Improve power by being smarter about sample-splitting (and when it needs to be done).

- R
Published by bdwilliamson almost 5 years ago

vimp - Enhanced estimation of cross-fitted AUC

Harmonize cross-fitted AUC estimation with cross-fitted deviance and R-squared estimation -- namely, that components of the estimator that don't involve regression (or machine learning) don't have to be estimated using cross-fitting (e.g., the variance of the outcome; or the probability that the outcome equals 1).

- R
Published by bdwilliamson almost 5 years ago

vimp - Allow IPW estimation in two-phase samples

Previously, only AIPW estimation was available for cases where the data arise from a two-phase sample; now IPW estimation is available as well.

- R
Published by bdwilliamson about 5 years ago

vimp - Bugfixes

• Update all tests to use glm rather than ranger or xgboost (improves package build and check speed and removes some dependencies)
• Internal bugfixes (mostly for cross-fitted VIM estimation)
• Update links and DOIs throughout vignettes and documentation

- R
Published by bdwilliamson about 5 years ago

vimp - Use package `stats` versions of internal functions, bugfixes

• Use stats::plogis and stats::qlogis for scale = "logistic"; this increases stability and R compatibility
• Minor bugfixes (e.g., if different number of observations are in each of the sample-splitting folds for hypothesis testing, there is no longer an error)

- R
Published by bdwilliamson over 5 years ago

vimp - Allow IPC weighting, CIs to be computed on non-identity scale

- R
Published by bdwilliamson over 5 years ago

vimp - Update inverse probability weighting behavior

- R
Published by bdwilliamson over 5 years ago

vimp - Add SPVIM

Add Shapley Population Variable Importance Measures (SPVIM) as an additional option for defining variable importance.

- R
Published by bdwilliamson over 5 years ago

vimp - Clean up and pass CRAN checks

Remove some unnecessary functions (e.g., cvvimnodonsker -- this has been superseded by cv_vim) and clean up \examples to pass CRAN checks.

- R
Published by bdwilliamson almost 6 years ago

vimp - Arbitrary predictiveness measures

This release allows you to define variable importance with respect to several different predictiveness measures: difference in classification accuracy, difference in AUC, and difference in deviance, in addition to difference in R-squared (this was previously the only way to measure variable importance supported by this package). Additionally, we now provide p-values for null hypothesis testing, and provide an updated interface for doing variable importance estimation (and cross-fitted variable importance estimation).

- R
Published by bdwilliamson almost 6 years ago

vimp - Fix CRAN errors

- R
Published by bdwilliamson over 6 years ago

vimp -

Fix dependency on SL.gam; update tests to cover more paths through code.

- R
Published by bdwilliamson over 6 years ago

vimp - v1.1.4

Bugfixes and performance updates for binary outcomes.

- R
Published by bdwilliamson over 7 years ago

vimp - v1.1.3.2

More small bugfixes; matches up to CRAN release 1.1.3

- R
Published by bdwilliamson over 7 years ago

vimp -

Small bugfix to 1.1.3, matches up code to CRAN release.

- R
Published by bdwilliamson over 7 years ago

vimp -

First GitHub stable release. Incorporates functions for variable importance and cross-validated variable importance.

- R
Published by bdwilliamson over 7 years ago