To maximize the speedup, we recommend always specifying a p -value threshold and using pulver as a filter to find models with significant or near-significant interaction terms.
← all excerpts
pulver: an R package for parallel ultra-rapid p-value computation for linear regression interaction terms.
1
—
—