Manifold_Dim

Higher is Better

[0, 1]

Description

Multi-method dimensionality efficiency score combining variance thresholds, Kaiser criterion, elbow detection, and spectral decay

mean(threshold_efficiencies, kaiser_efficiency, elbow_efficiency, decay_score)

Higher = more efficient compression (fewer dimensions capture variance). 1 = ideal dimensionality concentration.

Other metrics in the Intrinsic Latent Space (LSE) category

Rate of eigenvalue decay indicating information concentration in leading dimensions

Effective dimensionality measure from eigenvalue distribution (trajectory: lower is better, steady-state: higher is better)

Multi-method anisotropy combining log-ellipticity, condition numbers, ratio variance, entropy, dominance, and effective dimensionality