continuity

function continuity(cr, k): number;

Defined in: metrics/neighborhood.ts:157

Are the data’s neighbours still together in the projection?

The mirror of trustworthiness: it penalises points that were close in the data but got pushed apart — structure the projection hides rather than invents.

• Needs: high-dimensional data and projection. No labels.
• Range: [0, 1], higher is better; about 0.5 for a random projection.
• Cost: O(1), from an O(N² log N) pass.

k may not exceed maxKTrustworthiness(n) = floor(n / 2), as for trustworthiness — the two share a normaliser.

Parameters

Parameter	Type
cr	CoRanking
k	number

Returns

number

See

Venna & Kaski, Neural Networks 19 (2006) https://doi.org/10.1016/j.neunet.2006.05.014

Example

import { analyze, trustworthiness, continuity } from "@saehrimnir/sickle";

const a = analyze(data, projection);   // number[][] straight in

// Read both: they catch opposite failures and share one sweep.
trustworthiness(a.coRanking, 20);  // 0.9659 — invented neighbours
continuity(a.coRanking, 20);       // 0.9709 — hidden neighbours