Package index • mipdeval

Run iterative predictive analysis

All iterative predictive analyses begin with a call to run_eval(), supplying a either a PKPDsim model object or library and a NONMEM-style dataset.

run_eval(): Run iterative predictive analysis, looping over each individual's data
plot(<mipdeval_results>): Plot method for a run_eval() object

Set options for iterative predictive analysis

Pass these functions as arguments to run_eval() to set various options for iterative predictive analysis, including the control of how and whether various summary statistics are calculated.

fit_options(): Options for controlling MAP Bayesian fit
bootstrap_options(): Options for bootstrapping error metrics
stats_summ_options(): Options for summary statistics
vpc_options(): Options for VPC simulations

Calculate summary statistics for an iterative predictive analysis

These functions calculate various summary statistics from the results of run_eval(). All of these calculations can be included in the outputs of run_eval() by setting various options; they are primarly exported here for situations where additional post-processing is desired before running calculations.

calculate_bayesian_impact(): Calculate the impact of using Bayesian updating compared to population estimates
calculate_shrinkage(): Calculate eta-shrinkage
calculate_stats(): Calculate basic statistics, like RMSE, MPE, MAPE for forecasted data
calculate_bootstrap_summ(): Bootstrap confidence intervals for forecasting error metrics

Group observations in a NONMEM dataset

Helper functions for grouping observations in NONMEM datasets.

add_grouping_column(): Add grouping column using a function
group_by_dose(): Will create a separate group for each dose intervals that contains at least one sample
group_by_time(): Group data by time using bin separators

Built in datasets

nm_busulfan: Busulfan data
nm_vanco: Vancomycin data

Bootstrapping and error metrics

Lower-level functions for bootstrapping and calculating error metrics.

bootstrap_metrics() summarise_bootstrap_metrics() summarize_bootstrap_metrics(): Calculate metrics across bootstrapped folds
accuracy() is_accurate() is_accurate_abs() is_accurate_rel(): Accuracy
mape(): Mean absolute percentage error
mpe(): Mean percentage error
nrmse(): Normalized root-mean-squared error
rmse(): Root-mean-squared error

Compare results with PsN

Helper functions to compare the results of run_eval() with the PsN execute and proseval tools.

compare_psn_proseval_results() reldiff_psn_proseval_results() compare_psn_execute_results() reldiff_psn_execute_results(): Compare mipdeval results with PsN
parse_psn_proseval_results(): Parse PsN::proseval results.csv to filter out only the rows that we need (for prediction of next sample or group of samples)