Koopman Operators and Dynamic Mode Decomposition: Linearising Nonlinear Cash-Flow Swings and Forecasting Turning Points in Livestream Audience Curves: chapter answers
Livestream teams and shop operators widely believe two things: audience curves and daily takings are essentially a mess, and the safest way to read a turning point is a moving average. This book tests both beliefs with the Koopman operator and dynamic mode decomposition: curves that look chaotic are often sums of a few decaying or oscillating modes, and moving averages carry a built-in lag at turning points that comes from the structure of the filter. All computation uses the Python standard library alone, with singular value and eigen decompositions built from Jacobi rotations. The book's au…
DOI pending · monograph page goes live after publication · 5DL-B-238 · 中文版
- Chapter 1 · Two or Three Modes: the Low-Rank Structure of a Livestream Audience Curve and the Built-In Lag of Moving Averages
Chapter 1 shows that the first two modes of an illustrative audience curve hold 0.9285 of its energy, but a random walk also reaches 0.872, so explained variance is not evidence and out-of-sample forecasting is; a movin… - Chapter 2 · The Koopman Operator: Lifting Nonlinear Dynamics into Linear Ones, and the Limits of the Lift
Chapter 2 explains how a nonlinear system is lifted into a linear one, how a forced fit in raw coordinates gives a distorted eigenvalue of 0.744427, and why chaotic orbits have no finite set of modes. - Chapter 3 · Jacobi Rotations: Symmetric Eigendecomposition and the Singular Value Decomposition with the Standard Library Alone
A Jacobi rotation turns one off-diagonal element into zero, off-diagonal energy can only flow onto the diagonal, and cyclic sweeps converge quadratically in their later stage: in program p31 six sweeps take an 8 by 8 Gr… - Chapter 4 · Dynamic Mode Decomposition: Eigenvalues, Amplitudes and the False Decay Caused by Noise
Chapter 4 assembles dynamic mode decomposition and shows that noise makes a swing that never decays look like one with modulus 0.7744 and a half-life of 2.7 minutes, which a forward-backward correction repairs. - Chapter 5 · Lead Time at the Turning Point: a Mode-Based Warning Every Five Minutes against Moving Averages
Chapter 5 compares three rules over two hundred random illustrative sessions: the mode warning has a median lead of 3.45 minutes, comes on average 13.13 minutes before the moving average and is premature in 0.055 of ses… - Chapter 6 · Cash Flow: Linearising Daily, Weekly and Monthly Swings with Modes, and Choosing the Observable
Chapter 6 applies the method to a shop's daily cash flow, with a 28-day forecast error of 2301 yuan against 4665 for "the same day last week"; taking logarithms is a Koopman lift that cuts the rank-3 error from 0.1645 t… - Chapter 7 · East and West in the Sources: "Knowing the Incipient" in the Book of Changes and Fourier's Analysis of Heat
Chapter 7 examines "knowing the incipient" in the Book of Changes (ten minutes before the turn the second difference has a signal-to-noise ratio of only 0.031) and Fourier's analysis of heat (a fixed Fourier basis holds… - Chapter 8 · Closing the Empirical Loop: a Mode Fit Every Five Minutes, a Cooling Warning and a Pre-Registered Test
Chapter 8 gives the five-minute fitting procedure, the Online-Curve Turning-Point Early Warner and a pre-registered test plan: a lead difference of 3 minutes needs 51 sessions and the false-alarm test needs 68. Marked b…
Illustrative model calculation, not a live quote