For elliptical distributions, peeling the k smallest principal components maximizes total variance and Frobenius norm while peeling the k largest minimizes them, proving an unsupervised No Free Lunch theorem for bump-hunting.
The Futility of Bias-Free Learning and Search,
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
Conserved active information I^⊕ is introduced as a symmetric measure of net information change across an entire search space that respects No-Free-Lunch conservation and distinguishes disorder-increasing from order-imposing knowledge.
citing papers explorer
-
PRIM-cipal components analysis
For elliptical distributions, peeling the k smallest principal components maximizes total variance and Frobenius norm while peeling the k largest minimizes them, proving an unsupervised No Free Lunch theorem for bump-hunting.
-
Conserved active information
Conserved active information I^⊕ is introduced as a symmetric measure of net information change across an entire search space that respects No-Free-Lunch conservation and distinguishes disorder-increasing from order-imposing knowledge.