Perform Bayes Factor-Based Control for Estimating \(\pi_0\) — BF_control • fashr

This function estimates \(\pi_0\), the proportion of datasets that follow the null hypothesis, using Bayes Factor (BF) control.

BF_control(BF, plot = FALSE)

Arguments

BF: A numeric vector of Bayes Factors computed from `BF_compute()`.
plot: A logical value. If TRUE, generates diagnostic plots for BF control.

Value

A list containing:

mu: Cumulative mean of sorted Bayes Factors.
pi0_hat: Estimated \(\pi_0\) values for each BF threshold.
pi0_hat_star: Final estimated \(\pi_0\) based on the first BF threshold where \(E(BF) \geq 1\).

Examples

set.seed(1)
BF_values <- runif(100, 0.5, 5)  # Example Bayes Factors
BF_control_results <- BF_control(BF_values, plot = TRUE)

print(BF_control_results$pi0_hat_star)
#> [1] 0.15