A moving-window ridge-regression pipeline fits local linear operators, extracts a dominant 2D input-response plane by optimization or commutator methods, and tracks R = K/Kc(Δ), recovering this reduced non-normal geometry from finite multivariate time series with far fewer samples than full operator
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.data-an 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Inferring Non-Normal Amplification Geometry from Multivariate Time Series
A moving-window ridge-regression pipeline fits local linear operators, extracts a dominant 2D input-response plane by optimization or commutator methods, and tracks R = K/Kc(Δ), recovering this reduced non-normal geometry from finite multivariate time series with far fewer samples than full operator