Skip to contents

Unified outlier cleaning

One-call entry point — orchestrates the four primitives with the class-aware consensus rule (default; user-configurable via consensus =), plus topological block expansion. Two post-run diagnostics: mt_diagnose_clean_track() for cleaning health, mt_diagnose_flags() for per-detector flag composition + consensus-mode comparison.

mt_clean_track()
Unified outlier detection for a move2 track
mt_diagnose_clean_track()
Post-run diagnostic suite for mt_clean_track results
mt_diagnose_flags()
Per-detector flag audit for a cleaned move2 object
v_phys_estimate()
Allometric estimate of physiological maximum speed

Outlier-detection primitives

The four primitives composed by mt_clean_track(). Run individually for fine-grained control.

mt_flag_outliers_bridge()
Flag outliers via Brownian bridge residuals
mt_flag_outliers_detour()
Flag outliers using path-vs-displacement detour ratio
mt_flag_outliers()
Flag or remove outliers in movement data based on joint movement probabilities
mt_flag_speed_cap()
Flag fixes by a speed cap (data-driven or physiological)
mt_suggest_speed_cap()
Suggest a data-driven speed cap from the observed step-speed distribution
mt_peel_speed()
Iterative speed peel at a fixed physiological cap

Consensus rule (combine per-detector flags)

Externalised voting/consensus across detector outputs. Used internally by mt_clean_track() and exported so users running their own cascades (or running the four primitives standalone) can apply the same rules to their own per-detector flag columns.

mt_flag_consensus()
Combine per-detector flag columns into a single outlier decision

State-aware bridge variants (standalone)

Per-axis bridge motion variance with an envelope flag rule across three Z channels. Exported as standalone primitives, not currently wired into mt_clean_track()’s cascade.

mt_flag_outliers_dbgb()
Flag outliers via state-aware bridge residuals (dBGB envelope)
mt_flag_outliers_dbbmm()
Flag outliers via dBBMM-Z (isotropic submodel of dBGB)

Alternative outlier strategies

Voting / scanning alternatives on the probability surface. See mt_clean_track()’s @seealso for when each is the right reach.

mt_sequential_outliers()
Sequential outlier detection for movement data
mt_combined_outliers()
Combined outlier detection using multiple methods

Multi-scale persistence annotation

Detector-agnostic confidence helper that annotates flagged fixes with a persistence score across temporal scales.

mt_persistence_score()
Multi-scale persistence score for outlier flags

Pre-cleaning helpers

Filter fixes whose GPS geometry is unreliable before running the cleaner.

mt_filter_gps_quality()
Filter fixes by GPS-geometry quality

Utilisation distributions and motion variance

Dynamic Brownian-bridge variance estimation, UD computation on a common grid, accessors and segment masking for the variance pipeline.

mt_dbbmm_variance()
Dynamic Brownian Motion Variance Estimation
mt_dbbmm_ud()
Dynamic Brownian Bridge Movement Model — Utilisation Distribution
mt_dbgb_variance()
Dynamic Bivariate Gaussian Bridge Variance Estimation
mt_dbgb_ud()
Dynamic Bivariate Gaussian Bridge — Utilisation Distribution
mt_motion_variance()
Extract Motion Variance
mt_suggest_dbbmm_window()
Suggest dBBMM / dBGB sliding-window parameters from track sampling
ud_volume()
Convert a utilisation distribution to cumulative-volume space
ud_outer_probability()
Outer probability at a set of locations
mt_mask_segments()
Exclude segments from a dBBMM or dBGB utilisation distribution
emd()
Earth-mover's distance between utilisation distributions

Trajectory utilities

Track segmentation, thinning, corridor detection.

mt_corridor()
Identify corridor segments in a move2 trajectory
mt_thin_distance()
Thin a track by cumulative along-track distance
mt_thin_time()
Thin a track to a target time interval (tolerance-constrained)

S3 print methods

Internal print formatters for diagnostic objects.

print(<mt_dbbmm_variance>)
Print method for mt_dbbmm_variance
print(<mt_dbgb_variance>)
Print method for mt_dbgb_variance