Expand description
Online/streaming Φ estimation for time-series data.
Computes Φ over a sliding window of observations, maintaining an empirical TPM that is updated incrementally. Designed for:
- Neural data (EEG/fMRI time series)
- Real-time BCI applications
- Long-running consciousness monitoring
Key features:
- Exponential forgetting factor for non-stationarity
- Change-point detection in Φ trajectory
- EWMA smoothing for noise reduction
Structs§
- Streaming
PhiEstimator - Online Φ estimator with empirical TPM and sliding window.