Compute Asymmetry Scores for MONECA Segments — compute_asymmetry

Quantifies the degree of directional asymmetry within each segment at a given hierarchical level. For every within-segment pair (i, j), the asymmetry is `abs(RR_ij - RR_ji) / (RR_ij + RR_ji)`, ranging from 0 (perfectly symmetric) to 1 (completely one-directional).

Usage

compute_asymmetry_scores(segments, level = 2, method = "max_pair")

Arguments

segments: A moneca object returned by [moneca()] or [moneca_fast()].
level: Integer. Hierarchical level to evaluate (default 2).
method: Character. Aggregation method per segment: * `"max_pair"` (default): maximum pairwise asymmetry. * `"mean_pair"`: mean pairwise asymmetry weighted by the sum of the bidirectional relative risks (`rr_ij + rr_ji`). Note: this weights pairs by combined RR intensity, not by raw flow counts.

Value

A `data.frame` with one row per segment and columns:

segment_id: Integer segment index at the requested level.
n_members: Number of categories in the segment.
asymmetry_score: Aggregate asymmetry (max or mean).
max_pair_from: Origin category of the most asymmetric pair.
max_pair_to: Destination category of the most asymmetric pair.
max_pair_asym: Asymmetry value of the most asymmetric pair.

Details

The function retrieves the original mobility matrix from `segments$mat.list[[1]]` and computes an asymmetric relative risk matrix via [weight.matrix()] with `symmetric = FALSE` and `cut.off = 0`. Pairs with zero flow in both directions receive an asymmetry of 0. Pairs with flow in only one direction receive the maximum asymmetry of 1.

Examples

mob <- generate_mobility_data(n_classes = 6, seed = 42)
seg <- moneca(mob, segment.levels = 3)
#> 
  |                                                                            
  |                                                                      |   0%
  |                                                                            
  |============                                                          |  17%
#> 29% ready!
#> 
  |                                                                            
  |=======================                                               |  33%
  |                                                                            
  |===================================                                   |  50%
#> 57% ready!
#> 
  |                                                                            
  |===============================================                       |  67%
  |                                                                            
  |==========================================================            |  83%
#> 86% ready!
#> 
  |                                                                            
  |======================================================================| 100%
#> 100% ready!
#> 
#> 
  |                                                                            
  |                                                                      |   0%
#> 100% ready!
#> 
compute_asymmetry_scores(seg, level = 2)
#>   segment_id n_members asymmetry_score max_pair_from max_pair_to max_pair_asym
#> 1          1         3      0.23199448       Class 4     Class 6    0.23199448
#> 2          2         2      0.03168966       Class 2     Class 3    0.03168966