Interpretable Civic-Accountable and Responsible Machine Learning

[Up] [Top]

Documentation for package ‘civic.icarm’ version 0.3.0

Help Pages

civic_audit Generate a reproducible audit trail
civic_calibrate Assess probability calibration (binary classification)
civic_compare Compare multiple civic_models on a shared test set
civic_dashboard Generate a 4-panel civic accountability dashboard
civic_education civic_education dataset
civic_equalized_odds_curve Compute equalized odds curves across thresholds (binary only)
civic_equity_summary Summarise fairness into scalar equity indicators
civic_explain Generate global model explanations
civic_explain_local Generate local instance-level explanations
civic_fairness Compute group-level fairness metrics
civic_fit Fit an interpretable ICARM model — works with any tabular data
civic_german_credit civic_german_credit dataset
civic_metrics Compute performance metrics for any task type
civic_plots Visualisation functions for civic.icarm
civic_plot_calibration Plot calibration curve
civic_plot_comparison Plot multi-model comparison
civic_plot_confusion Plot confusion matrix
civic_plot_fairness Plot group-level fairness metric
civic_plot_importance Plot feature importance
civic_plot_roc_groups Plot per-group ROC curves
civic_plot_thresholds Plot threshold performance curves
civic_racism_survey Synthetic Racism and Civic Participation Survey
civic_scorecard Generate a full civic accountability scorecard
civic_split Reproducible train/test split
civic_thresholds Threshold sweep for binary classification
civic_voting civic_voting dataset
predict.civic_model Predict from a civic_model
print.civic_model Print a civic_model
summary.civic_model Summary of a civic_model