Kernel Profile Calculator

English version for psychometrics.eth.limo. Implements a Gaussian-kernel similarity between an observation x and a reference profile m with optional per-feature weights and bandwidth (σ). Equation (6): K(x,m) = exp( - Σ wᵢ (xᵢ − mᵢ)² / (2 σ²) )

Manual input

Observation vector x (comma-separated)

Reference vector m (comma-separated)

Feature weights w (comma-separated; optional)

Bandwidth σ (positive)

Batch via CSV

Upload a CSV with columns x1..xN for observations and a single row for the reference m1..mN (or provide m above). Optional columns w1..wN for weights.

Visualization

Scores K(x,m) in [0,1]. Higher = closer to reference profile.

Notes

All vectors must have the same length N. Missing weights default to 1.
Numerical stability: if σ≤0 or any non-numeric value is found, you’ll get an error.
This page is static (client‑side); safe to host on IPFS/eth.limo.