Attractor Reconstruction

Correlation dimension, Lyapunov exponent, attractor filling
dynamicaldim phase space7 metrics

What It Measures

The dimension and divergence rate of the signal's phase-space attractor.

Delay-embeds the time series (at dimensions 2 through 8 for correlation dimension, up to 10 for Lyapunov exponent) using the first zero-crossing of the autocorrelation as the lag. In this reconstructed space, applies Grassberger-Procaccia to estimate the correlation dimension D2 (how many dimensions the attractor fills) and Rosenstein's method to estimate the maximum Lyapunov exponent (how fast nearby trajectories diverge).

Metrics

correlation_dimension

How many effective dimensions does the attractor fill? Collatz Stopping Times leads at 4.22: its complex branching dynamics fill a roughly 4D manifold. Neural Net Dense (4.02) and ECG Supraventricular (4.01) are similarly high-dimensional. The Lorenz attractor sits around 2.05 (textbook D2 for the Lorenz system). Constants and Fibonacci Word score 0.0 — degenerate point or 1D attractors.

d2_saturation

Does the dimension estimate converge as you increase the embedding dimension? Champernowne (0.997) and Triangle Wave (0.997) saturate immediately — their low intrinsic dimension is captured at the lowest embedding. Collatz Parity scores 0.0 (dimension never converges, suggesting the signal doesn't live on a finite-dimensional manifold). High saturation means you can trust the D2 estimate; low saturation means the attractor is higher-dimensional than the embedding can capture.

filling_ratio

What fraction of the embedding space does the trajectory actually visit? Dice Rolls (0.994) and XorShift32 (0.982) fill almost all of it — they're space-filling in delay coordinates. Logistic Period-2 scores 0.002 (the trajectory visits only two points in any embedding). This separates low-dimensional attractors from space-filling noise.

lyapunov_max

The maximum Lyapunov exponent: how fast do nearby trajectories diverge? Positive means chaos (exponential separation), zero means periodic or quasiperiodic, negative means contracting. Henon Near-Crisis leads at 0.106: it's on the edge of destruction, with maximum divergence. Financial returns (Nikkei -0.003, NYSE -0.0003) are slightly negative — they're mean-reverting on short timescales.

lyap_sum

Sum of all positive Lyapunov exponents in the reconstructed spectrum (Wolf-style multi-direction estimate, capped at the embedding dimension). Positive total = expanding directions exceed contracting ones (dissipative chaos). Anderson 1D Localized (+4.20) and Langton's Ant (+4.18) lead; Aubry-André Critical and ECG Beat Conformity follow. Strongly bimodal: ~64/298 sources NaN where the spectrum is degenerate (constants, near-periodic, alphabet-bound symbolic). When valid, separates "many directions expand a little" from "one direction expands a lot."

lyap_entropy

Kolmogorov-Sinai entropy estimate as the sum of positive exponents from the same spectrum. Fibonacci Tight-Binding (5.00), Langton's Ant (4.60), and Anderson 1D Localized (4.60) saturate the cap — their reconstructed dynamics produce many expanding directions simultaneously. Distinct from lyapunov_max (which sees only the steepest exponent) and from Information Theory's entropy (which sees only the marginal distribution).

kaplan_yorke_dim

Kaplan-Yorke / Lyapunov dimension: a fractional estimate of the attractor's information dimension from the spectrum. MFPT Outer Race / Shuffled Blocks (6.00) saturate at the embedding ceiling; structured low-dimensional dynamics (logistic period orbits, Fibonacci Word) collapse to 0.0. Complements correlation_dimension: D2 measures attractor geometry directly; D_KY infers it from divergence rates, and the two disagree where finite-sample bias hits one but not the other.

Atlas Rankings

d2_saturation
SourceOriginValue
Logistic r=3.5 (Period-4)iterated-maps0.9974
Damped Pendulumcontinuous-flows0.9972
Triangle Wavesignal-synthesis0.9969
···
Thue-Morsesymbolic-dynamics0.0000
Fibonacci Wordsymbolic-dynamics0.0000
Periodic Wordsymbolic-dynamics0.0000
embedding_divergence_sum
SourceOriginValue
Anderson 1D Localizedrandom-matrix-quantum4.2021
Langton's Antsymbolic-dynamics4.1824
Aubry-André Criticalrandom-matrix-quantum3.5017
···
Sandpileself-organized-criticality-19.6113
ECG Fusionphysiological-19.4862
Collatz Flightsnumber-theory-19.4048
embedding_spread_dimension
SourceOriginValue
Gaussian Noisestochastic-process4.2978
Network Packet Sizesalgorithmic-bytes4.2382
Shuffled Blocksstochastic-process4.2177
···
Pell Wordsymbolic-dynamics0.0000
Collatz Cycle Wordnumber-theory0.0000
Golden-Mean β-Shiftsymbolic-dynamics0.0000
filling_ratio
SourceOriginValue
Dice Rollsstochastic-process0.9938
Euler-Mascheroni γ Digitsnumber-theory0.9830
ChaCha20algorithmic-bytes0.9829
···
Logistic r=3.2 (Period-2)iterated-maps0.0020
Logistic r=3.5 (Period-4)iterated-maps0.0020
Logistic r=3.83 (Period-3 Window)iterated-maps0.0029
kaplan_yorke_dim
SourceOriginValue
MFPT Outer Racemachine-vibration6.0000
Shuffled Blocksstochastic-process6.0000
MFPT Normalmachine-vibration5.9557
···
Triangle Wavesignal-synthesis0.0000
Sine Map (Feigenbaum)iterated-maps0.0000
μ-law Sinesignal-synthesis0.2013
lyap_entropy
SourceOriginValue
Fibonacci Tight-Bindingrandom-matrix-quantum5.0000
Langton's Antsymbolic-dynamics4.6816
Anderson 1D Localizedrandom-matrix-quantum4.5984
···
Triangle Wavesignal-synthesis0.0000
Sine Map (Feigenbaum)iterated-maps0.0000
Sine Wavesignal-synthesis0.0001
lyapunov_max
SourceOriginValue
Fibonacci Tight-Bindingrandom-matrix-quantum0.8698
Tent Mapiterated-maps0.5712
von Mangoldt Functionnumber-theory0.5361
···
macOS Mach-O (dyld)algorithmic-bytes0.0000
DNA Phage Lambdagenomic0.0000
Bzip2 (level 1)algorithmic-bytes0.0000

When It Lights Up

Attractor Reconstruction provides the classic chaos diagnostic: positive Lyapunov with finite correlation dimension means deterministic chaos. The framework uses it alongside Gottwald-Melbourne (which doesn't need embedding) as a cross-check. In the atlas, correlation_dimension separates the dynamical view's low-dimensional chaos cluster (D2 = 2-4: Lorenz, Rossler, Henon) from noise (D2 saturates at embedding dimension) and periodicity (D2 = 1). The Lyapunov-spectrum triple (lyap_sum / lyap_entropy / kaplan_yorke_dim) adds a dissipativity axis the single max-Lyapunov can't see — useful where many directions expand weakly versus one direction expanding strongly.

Open in Atlas
← Navier-StokesBoltzmann →