R/compute_nulls.R

In StoreyLab/lfa: Logistic Factor Analysis for Categorical Data

.compute_nulls <- function(P, d, B, max_iter = 100, tol = 1e-10) {
    m <- nrow(P)
    n <- ncol(P)

    # since m and n are integers, multiplying them causes a buffer overflow
    # let's multiply them as doubles, overcomes the problem
    n_data <- (n + 0) * (m + 0)

    stats0 <- matrix(0, m, B)
    for (i in seq_len(B)) {
        X0 <- matrix(stats::rbinom(n_data, 2, P), nrow=m, ncol=n)
        LF0 <- lfa(X0, d)
        # this calculates stats correctly, even when X0 is BEDMatrix!
        stats0[, i] <- .gof_stat(X0, LF0, max_iter=max_iter, tol=tol)
    }

    return(stats0)
}