Skip to contents

Estimate angular marginal support

Source: R/directional-uncertainty.R
marginal_support_data.Rd

Computes a kernel-weighted marginal support score along the conditioning angle. Low relative support marks angular regions where conditional displays should be interpreted cautiously.

Usage

marginal_support_data(
  theta,
  grid = NULL,
  n_theta = 181,
  kappa = 20,
  relative_threshold = 0.15
)

Arguments

theta

Conditioning angles in radians.

grid

Optional angular grid in radians.

n_theta

Number of grid points used when grid is NULL.

kappa

Circular concentration for local kernel weights.

relative_threshold

Relative support threshold used to flag low-support regions. With other ingredients fixed, the local standard error is approximately inflated by 1 / sqrt(relative_support). The default 0.15 corresponds to an inflation of about 2.58 relative to the best-supported direction; 0.25 is a more conservative threshold corresponding to about twofold inflation.

Value

A data frame with grid values, support, relative support and flags.

See also

Other directional uncertainty helpers: autoplot.bootstrap_stat_orbit(), autoplot.bootstrap_toroidal_ridge(), bootstrap_stat_orbit(), bootstrap_toroidal_ridge(), plot_marginal_support(), sensitivity_bandwidth_grid(), simulate_independence_lineup()

Examples

dat <- simulate_cylindrical(n = 60, seed = 1)
marginal_support_data(dat$theta, n_theta = 12)
#>        theta  support relative_support low_support kappa relative_threshold
#> 1  0.2617994 4.117877        0.5616172       FALSE    20               0.15
#> 2  0.7853982 4.047707        0.5520471       FALSE    20               0.15
#> 3  1.3089969 5.720164        0.7801454       FALSE    20               0.15
#> 4  1.8325957 3.627064        0.4946777       FALSE    20               0.15
#> 5  2.3561945 7.332177        1.0000000       FALSE    20               0.15
#> 6  2.8797933 6.249973        0.8524034       FALSE    20               0.15
#> 7  3.4033920 5.291841        0.7217285       FALSE    20               0.15
#> 8  3.9269908 6.008550        0.8194770       FALSE    20               0.15
#> 9  4.4505896 7.067453        0.9638957       FALSE    20               0.15
#> 10 4.9741884 6.890080        0.9397045       FALSE    20               0.15
#> 11 5.4977871 4.728663        0.6449194       FALSE    20               0.15
#> 12 6.0213859 3.597078        0.4905880       FALSE    20               0.15