Skip to contents

Compare two networks descriptively

Source: R/compare_network.R
compare_model.Rd

Computes a battery of descriptive comparison metrics between two networks or two weight matrices: weight deviations (mean / median / RMS / max absolute difference, relative mean absolute difference, coefficient-of- variation ratio), four correlation measures (Pearson, Spearman, Kendall, distance correlation), five dissimilarity measures (Euclidean, Manhattan, Canberra, Bray-Curtis, Frobenius), five similarity measures (Cosine, Jaccard, Dice, Overlap, RV), pattern agreements, and side-by-side network metrics. Optionally adds centrality differences and centrality correlations.

Usage

compare_model(x, ...)

# S3 method for class 'netobject'
compare_model(
  x,
  y,
  scaling = "none",
  measures = character(0),
  network = TRUE,
  ...
)

# S3 method for class 'cograph_network'
compare_model(
  x,
  y,
  scaling = "none",
  measures = character(0),
  network = TRUE,
  ...
)

# S3 method for class 'matrix'
compare_model(
  x,
  y,
  scaling = "none",
  measures = character(0),
  network = TRUE,
  ...
)

Arguments

x

A netobject, cograph_network, or numeric square matrix.

...

Ignored.

y

A netobject, cograph_network, or numeric square matrix.

scaling

Scaling applied to both weight matrices before comparison. One of:

"none"

Identity (default).

"minmax"

\((w - \min) / (\max - \min)\); maps to \([0, 1]\).

"max"

\(w / \max(|w|)\); preserves sign.

"rank"

Min-max of average ranks; ordinal scaling.

"zscore"

\((w - \bar w) / s_w\); standard score.

"robust"

\((w - \mathrm{med}(w)) / \mathrm{mad}(w)\); Huber-style robust z-score, resists outliers.

"log"

\(\log(w)\); requires \(w > 0\).

"log1p"

\(\log(1 + w)\); admits \(w \ge 0\).

"softmax"

Numerically stable softmax over the flattened vector.

"quantile"

Empirical CDF of the flattened vector.

"frobenius"

Divide the matrix by its Frobenius norm \(\|W\|_F = \sqrt{\sum w_{ij}^2}\); matrix-level normalisation.

"row"

Row-stochastic normalisation (each row's absolute values sum to 1). Only meaningful for non-negative matrices; rows summing to zero are left unchanged.

Scalings that produce negative weights (zscore, robust) are compatible with network = TRUE because the side-by-side metrics use Nestimate's base-R Floyd-Warshall, which handles negative weights.

measures

Character vector of centrality measures to compare. Empty by default (no centrality block). Valid names are "InStrength", "OutStrength", and "Betweenness". Unknown names are ignored with a warning.

network

Logical. Include side-by-side network metrics from summary()? Default TRUE.

Value

A net_comparison object: a named list with matrices, difference_matrix, edge_metrics, summary_metrics, optionally network_metrics, centrality_differences, centrality_correlations.

Details

Mirrors tna::compare() numerically. Inputs are converted to weight matrices and scaled before comparison; the choice of scaling determines how weights from different estimators are placed on a common footing.