Limitations
SheShe provides tools for modal analysis, but several limitations apply:
Dependence on surface representation
The algorithms assume a smooth and well-sampled surface. Imperfect meshes or sparse points degrade boundary estimates.
Mitigation: preprocess meshes to remove noise, interpolate missing regions and validate against known benchmarks.
High-dimensional approximations
In very high dimensions SheShe relies on approximations that may miss subtle structures and increase computation time.
Mitigation: apply dimensionality reduction or focus on informative subspaces with SubspaceScout.
Noise and sampling sensitivity
Noisy or sparsely sampled data can produce unstable boundaries.
Mitigation: denoise inputs and aggregate multiple runs through ModalScoutEnsemble.
Parameter tuning
Performance depends on choosing suitable bandwidths and clustering thresholds.
Mitigation: explore settings via the Quick API and cross-validation.