hampel

Sequential outlier detection and removal using Hampel identifiers.

Supports f32 and f64. Works without std — no heap allocation required.

What is a Hampel Identifier?

A Hampel identifier is a robust method for detecting outliers in sequential (streaming) data. It maintains a sliding window over the most recent N values and checks each incoming point against a threshold derived from the median and MAD (Median Absolute Deviation) of the window.

The key advantage over mean-based methods is robustness: even if the window already contains a few outliers, the median is unaffected and the estimate stays accurate.

When an outlier is detected, the value is replaced — either with the window median (default) or a linearly extrapolated value (see the extrapolation feature below).

Usage

Add this to your Cargo.toml:

[dependencies]
hampel = "0.2"
# features = ["extrapolation"]  # optional — see below

extrapolation feature

When this feature is enabled, linear extrapolated values are returned when outliers are detected. When disabled (default), the median value of the window is returned.

Example

use hampel::Window;

fn main() {
    // Window size : 5  (must be >= 3)
    // Init value  : 0.0
    // Threshold k : 3.0  →  flag if |x − median| > 3 × σ̂
    let mut filter = Window::<f64, 5>::new(0.0, 3.0);

    let input_vals = [0.0; 100];   // ← replace with your data (may contain outliers)
    let mut filtered_vals = [0.0; 100];

    for (i, val) in input_vals.iter().enumerate() {
        filtered_vals[i] = filter.update(*val);
    }
    // filtered_vals <-- Outliers have been removed
}

Sample output

Default (median replacement)

With extrapolation feature

License

Licensed under either of Apache License, Version 2.0 or MIT License at your option.

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.