Skip to contents

Direction

Functions for estimating private principal component directions.

dp_pc_dir()
Estimate principal component directions

Scree

Functions for estimating and visualizing differentially private scree values. The control helpers specify method-specific tuning options for the clipped, private modified winsorized mean, and Huber-type scree estimators.

dp_scree()
Differentially private scree values
dp_scree_plot()
Plot differentially private scree estimates
clipped_control()
Control options for clipped scree estimation
pmwm_control()
Control options for private modified winsorized scree estimation
huber_control()
Control options for Huber scree estimation

Score

Functions for constructing differentially private PCA score summaries, score plots, and group-wise score visualizations.

dp_score()
Differentially private score histograms
dp_score_plot()
Plot differentially private score histograms
dp_score_group()
Group-wise differentially private score histograms
dp_score_plot_group()
Plot group-wise differentially private score histograms

shiny app

shiny app for exploring differentially private PCA scree and score plots through a graphical interface.

dppca_app()
Launch the dppca Shiny app

Data

Datasets included for examples and demonstrations.

adult
Adult numeric data
gau
Five Gaussian clusters
gau_g
Five Gaussian clusters with group labels