Within-sample standardization turns Random Fourier Feature kernels into training-set dependent kernels, and calibrated minimax bounds suggest typical financial samples cannot support genuine high-dimensional learning.
The crucial distance term remains ∥fj−fℓ∥2 L2(µ) because the test error is measured with respect to the population distribution of x
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
q-fin.ST 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
High-Dimensional Learning in Finance
Within-sample standardization turns Random Fourier Feature kernels into training-set dependent kernels, and calibrated minimax bounds suggest typical financial samples cannot support genuine high-dimensional learning.