Guides¶
Use the guides when you know the scientific task but need help choosing the representation, estimator, diagnostics, uncertainty procedure, or reporting contract.
Data & representation¶
Start here when the main issue is how gaze should be represented before modeling.
- Preparing trajectories
- Native irregular trajectories
- Sparse irregular / PACE FPCA
- Audited native sparse-FPCA bandwidth selection — opt-in tuning with explicit measurement-noise limits
- Sparse planar MFPCA / joint PACE
- Asynchronous sparse planar MFPCA / joint PACE
- Observation-process diagnostics
- Basis representations
- Derived continuous functions
FPCA, uncertainty & review¶
Use these guides for functional decomposition, component retention, stability, uncertainty, anomaly review, participant-aware resampling, and prediction from partially observed sparse histories.
- FPCA & MFPCA
- FPCA stability & validation
- Selecting FPC count
- Predictive FPC selection
- FPC shape uncertainty
- Simultaneous FPC bands
- FPCA spectrum uncertainty
- FPC score basis uncertainty
- Sparse partial-trajectory prediction
- Simultaneous functional mean bands
- Near-tied FPC subspaces
- Outliers & influence
- Conformal anomaly review
Registration, multilevel & composition¶
Use these when timing deformation, participant/trial hierarchy, or AOI composition is itself part of the scientific object.
- Registration & phase
- Phase FPCA & registration sensitivity
- Multilevel FPCA
- AOI probability functions
- Elastic trajectory analysis
Functional regression & repeated measures¶
Use these for experimental predictors, functional outcomes, repeated trials, generalized outcomes, and FPCR uncertainty.
- Functional regression & clustering
- Function-on-scalar regression
- Functional mixed-effects regression
- Gaussian FPCR bootstrap uncertainty
- Heteroscedastic FPCR wild bootstrap
The navigation expands this category into the full set of generalized, mixed-effects, simultaneous-inference, covariance, and prediction guides.
Geometry & nonlinear dynamics¶
Use these when the research question concerns trajectory shape, ordered-path distance, recurrence, nonlinear structure, Lyapunov behavior, or directed information flow.