Package index • circlss

Model fitting

Fit circular-response GAM, the classical (parametric) circular regressions

circ_gam() predict(<circ_gam>) fitted(<circ_gam>) print(<circ_gam>): Circular-response GAM
circ_lm() coef(<circ_lm>) fitted(<circ_lm>) residuals(<circ_lm>) logLik(<circ_lm>) predict(<circ_lm>) print(<circ_lm>): Classical circular regression
plot(<circ_gam>): Plot a circular-response GAM fit
plot(<circ_lm>): Plot a classical circular regression fit

Diagnostics

Circular residuals and the diagnostic panel grid for circ_gam() and circ_lm() fits – the circular analogue of plot.lm() and gam.check(), complementing the effect-display plot methods.

circ_resid(): Circular regression residuals
circ_check(): Diagnostic panels for a circular regression fit

Mixture models

An EM engine for finite mixtures of circular distributional GAMs – density clustering and the circular-linear / circular-circular / linear-circular regression trio.

circ_mix() circ_mix.control() print(<circ_mix>) summary(<circ_mix>) logLik(<circ_mix>) coef(<circ_mix>) predict(<circ_mix>): Finite mixtures of circular distributional GAMs by EM
plot(<circ_mix>): Plot a fitted circular mixture
circ_kmeans(): Circular k-means clustering

Response families

mgcv general families for angular responses: one linear predictor — and so one formula, with smooths — per distribution parameter.

vmlss(): von Mises location-scale family
pnlss(): Projected normal location family
wclss(): Wrapped Cauchy location-scale family
wnlss(): Wrapped normal location-scale family
cardlss(): Cardioid location-scale family
cartlss(): Cartwright power-of-cosine location-scale family
jplss(): Jones-Pewsey location-concentration-shape family
ssjplss(): Sine-skewed Jones-Pewsey location-concentration-shape-skewness family
kjlss(): Kato-Jones Mobius four-parameter family
vmftlss(): Flat-topped von Mises location-concentration-shape family
ibslss(): Inverse Batschelet location-concentration-skewness-peakedness family
ajplss(): Asymmetric Jones-Pewsey location-concentration-shape-asymmetry family

Linear-response families

Weight-aware location-scale families for a linear response with circular covariates (l~c), drop-in compatible with circ_gam().

gausslss(): Weight-aware Gaussian location-scale family
gammalss(): Weight-aware gamma location-scale family

Utilities

Angle conversion and wrapping helpers (radians/degrees, branch wrapping).

rad() deg() wrap(): Radian conversion helpers

Datasets

Circular-response data sets worked through the package: the classical (circ_lm), additive multi-covariate (circ_gam), and bimodal mixture (circ_mix) examples.

periwinkles: Movements of small blue periwinkles
songbirds: Flight orientation of nocturnal migrating songbirds
windfarm: Wind direction at a South African wind farm (ten-minute records)