Package: HMC 1.2

Tianyu Zhang

HMC: High-Dimensional Mean Comparison with Projection and Cross-Fitting

Provides interpretable high-dimensional mean comparison methods (HMC). For example, users can apply these methods to assess the difference in gene expression between two treatment groups. It is not a gene-by-gene comparison. Instead, the methods focus on the interplay between features and identify those that are predictive of the group label. The tests are valid frequentist procedures and yield sparse estimates indicating which features contribute to the group differences.

Authors:Tianyu Zhang [aut, cre, cph]

HMC_1.2.tar.gz
HMC_1.2.tar.gz(r-4.7-any)HMC_1.2.tar.gz(r-4.6-any)
HMC_1.2.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
card.svg |card.png
HMC/json (API)

# Install 'HMC' in R:
install.packages('HMC', repos = c('https://cran.r-universe.dev', 'https://cloud.r-project.org'))

Bug tracker:https://github.com/terrytianyuzhang/hmc/issues

On CRAN:

Conda:

1.61 score 41 scripts 562 downloads 24 exports 14 dependencies

Last updated from:ed3faa5247. Checks:4 OK. Indexed: yes.

TargetResultTimeFilesSyslog
linux-devel-x86_64OK130
source / vignettesOK176
linux-release-x86_64OK141
wasm-releaseOK97

Exports:anchored_lasso_testingcheck_data_for_foldscheck_non_null_and_identical_colnamescollect_active_features_projcombine_folds_mean_diffcompute_predictive_contributionsdebiased_pc_testingestimate_leading_pcestimate_nuisance_parameter_lassoestimate_nuisance_pcevaluate_influence_function_multi_factorevaluate_pca_lasso_plug_inevaluate_pca_plug_inextract_lasso_coefextract_pcfit_lassoindex_splitermean_comparison_anchornormalize_and_splitprocess_fold_mean_diffsimple_pc_testingsummarize_feature_namesummarize_pc_namevalidate_and_convert_data

Dependencies:codetoolsforeachglmnetgrpregirlbaiteratorslatticeMASSMatrixPMARcppRcppEigenshapesurvival

Readme and manuals

Help Manual

Help pageTopics
Anchored test for two-sample mean comparison.anchored_lasso_testing
Check that data has enough rows for cross-validation foldscheck_data_for_folds
Check non-null and consistent column names across datasetscheck_non_null_and_identical_colnames
Collect active features and groups based on projection directionscollect_active_features_proj
Combine fold-level test statistics from cross-validationcombine_folds_mean_diff
Compute predictive contributions of feature groupscompute_predictive_contributions
Debiased one-step test for two-sample mean comparison. A small p-value tells us not only there is difference in the mean vectors, but can also indicates which principle component the difference aligns with.debiased_pc_testing
Estimate the leading principal componentestimate_leading_pc
The function for nuisance parameter estimation in anchored_lasso_testing().estimate_nuisance_parameter_lasso
The function for nuisance parameter estimation in simple_pc_testing() and debiased_pc_testing().estimate_nuisance_pc
Calculate the test statistics on the left-out samples. Called in debiased_pc_testing().evaluate_influence_function_multi_factor
Calculate the test statistics on the left-out samples. Called in anchored_lasso_testing().evaluate_pca_lasso_plug_in
Calculate the test statistics on the left-out samples. Called in simple_pc_testing().evaluate_pca_plug_in
Extract the lasso estimate from the output of anchored_lasso_testing().extract_lasso_coef
Extract the principle components from the output of simple_pc_testing() and debiased_pc_testing().extract_pc
Fit a (group) Lasso logistic regression classifierfit_lasso
Split indices into foldsindex_spliter
High-dimensional two-sample mean comparison with anchored projectionmean_comparison_anchor
Normalize and split two datasets using pooled mean and standard deviationnormalize_and_split
Process one cross-validation fold for mean difference testingprocess_fold_mean_diff
Simple plug-in test for two-sample mean comparison.simple_pc_testing
Summarize the features (e.g. genes) that contribute to the test result, i.e. those features consistently show up in Lasso vectors.summarize_feature_name
Summarize the features (e.g. genes) that contribute to the test result, i.e. those features consistently show up in the sparse principle components.summarize_pc_name
Validate and convert input datavalidate_and_convert_data