Recurrence networks and RQA

Primary reference: Marwan et al. (2007) Phys. Rep. 438, 237–379.

Recurrence networks connect embedded phase-space points when their distance falls below threshold ε. Recurrence quantification analysis (RQA) summarizes the recurrence matrix (RR, DET, LAM, etc.).

Assumptions

• Embedding dimension m and delay tau affect topology; defaults are chosen for univariate smoke tests.

• Larger ε yields denser graphs (fixture: recurrence_epsilon_monotonicity).

• Periodic signals yield high determinism DET (fixture: rqa_sine_determinism).

API

• ts2net.core.recurrence.RecurrenceNetwork

• ts2net.graphs.recurrence.recurrence_quantification()

• ts2net.stats.threshold_sensitivity.threshold_sensitivity_sweep()