mrreFalse

function mrreFalse(cr, k): number;

Defined in: metrics/neighborhood.ts:510

Rank error among the neighbours the projection shows.

Like trustworthiness, but graded: a neighbour displaced by one rank costs far less than one displaced by a hundred, rather than both simply counting as wrong.

• Needs: high-dimensional data and projection. No labels.
• Range: [0, 1], higher is better.
• Cost: O(k), from an O(N² log N) pass.

Parameters

Parameter	Type
cr	CoRanking
k	number

Returns

number

See

Lee & Verleysen, Neurocomputing 72 (2009) https://doi.org/10.1016/j.neucom.2008.12.017

Example

import { analyze, mrreFalse, mrreMissing } from "@saehrimnir/sickle";

const a = analyze(data, projection);

// Rank-weighted siblings of trustworthiness and continuity.
mrreFalse(a.coRanking, 20);    // 0.9232 — neighbours the projection invented
mrreMissing(a.coRanking, 20);  // 0.9400 — neighbours it lost