Function intersect_error_bound 

pub fn intersect_error_bound(
    a: &HyperLogLog,
    b: &HyperLogLog,
) -> Result<f64, HllError>

Absolute error the inclusion-exclusion estimate carries at one standard deviation. It scales with |A| + |B|, so a thin overlap between two large sets can come back with an error bar wider than the answer. Check it against the estimate before believing an intersection.

Examples found in repository

examples/sample_app.rs (line 190)

fn cross_venue_overlap(tape: &[Event]) {
    use subms_hyperloglog::{estimate_intersect, estimate_union, intersect_error_bound};
    println!("\n== venues: account reach and overlap ==");

    let mut a = HyperLogLog::new(14);
    let mut b = HyperLogLog::new(14);
    for e in tape {
        if e.venue == 0 {
            a.add_u64(e.account);
        } else {
            b.add_u64(e.account);
        }
    }
    let union = estimate_union(&a, &b).expect("same precision");
    let inter = estimate_intersect(&a, &b).expect("same precision");
    let bound = intersect_error_bound(&a, &b).expect("same precision");
    println!("  venue 0: {:>7.0} accounts", a.estimate());
    println!("  venue 1: {:>7.0} accounts", b.estimate());
    println!("  reach:   {union:>7.0} (true 50000)");
    println!("  both:    {inter:>7.0} (true 10000) +/- {bound:.0}");
    assert!(
        (union - 50_000.0).abs() / 50_000.0 < 0.05,
        "reach within 5%, got {union}"
    );
    assert!(inter > 0.0, "a 10k overlap must survive the subtraction");
}