Package 'designpower'

Find Optimal Design

Description

Iteratively search for optimal design parameters using adaptive simulation.

Usage

find_power(
  design,
  design_digits,
  opti,
  sim_power,
  extra_args = list(),
  power = 0.9,
  alpha = 0.1,
  filename = "power.duckdb"
)

Arguments

design	List. Initial design parameters.
design_digits	Named numeric. Precision (number of decimal places) for each parameter. The names must match those in design.
opti	Character. Name of parameter to optimize.
sim_power	Function. Simulation function (should accept design and n_sim).
extra_args	List. Optional additional arguments passed to sim_power. Defaults to an empty list.
power	Numeric. Target power (default 0.9).
alpha	Numeric. Significance level (default 0.1).
filename	Character. Path to DuckDB database file.

Details

Uses DuckDB to store simulations and gradually refines design parameters to achieve target power.

Value

Numeric vector. The optimized parameter value and confidence range.

Examples

## Not run: 
find_power(
  design = list(trend = -0.03, n_site = 20, n_year = 12),
  design_digits = c(trend = 4, n_site = 0, n_year = 0),
  opti = "trend",
  sim_power = sim_power,
  extra_args = list()
)

## End(Not run)

---

Simple Power Simulation

Description

A basic example of a power simulation function.

Usage

sim_power(
  design = list(trend = -0.03, n_year = 12, n_sample = 20),
  n_sim = 100,
  intercept = 2,
  sigma_error = 0.1
)

Arguments

design	A named list with design parameters: trend: The true trend parameter n_year: Number of years n_sample: Number of samples per year
n_sim	Integer. Number of simulations.
intercept	Numeric. Intercept of the model.
sigma_error	Numeric. Standard deviation of the error term.

Details

Simulates data under a simple model and returns p-values.

Value

A data frame with columns:

• p: P-values from simulations

• trend: Estimated trends

Examples

## Not run: 
sim_power(
  design = list(trend = -0.03, n_year = 12, n_sample = 20),
  n_sim = 10
)

## End(Not run)

---