get_edge_weights: Calculate edge weights

In mappeR: Construct and Visualize TDA Mapper Graphs

Calculate edge weights

Description

Calculate edge weights

Usage

get_edge_weights(overlap_lengths, cluster_sizes, edges)

Arguments

overlap_lengths	A named vector of cluster overlap lengths, obtained by calling length() on the output from [get_overlaps()].
cluster_sizes	A vector of cluster sizes.
edges	A 2D array of source and target nodes, representing an edge list. Should be ordered consistently with the overlap_lengths parameter.

Details

This value is calculated per edge by dividing the number of data points in the union of the two clusters by the number of data points in the intersection. Formally,

w(\{c_i, c_j\}) = \dfrac{|c_i \cap c_j|}{|c_i \cup c_j|} = \dfrac{|c_i \cap c_j|}{|c_i| + |c_j| - |c_i \cap c_j|}

Value

A vector of real numbers representing the Jaccard index of each overlap.

mappeR documentation built on June 9, 2025, 5:08 p.m.