Calculate the Wang similarity matrix

This function calculates the Wang similarity, based on the DAG for a given ontology. This function will return the full similarity matrix.

Usage

calculate_wang_sim_mat(parent_child_dt, weights)

Arguments

parent_child_dt

data.table. The data.table with column parent and child. You also need to have a type column for the Wang similarity to provide the weights for the relationships.

weights

Named numeric. The relationship of type to weight for this specific edge. For example c("part_of" = 0.8, "is_a" = 0.6).

Value

The symmetric Wang similarity matrix.

Examples

# Wang similarity matrix with relationship-specific weights
onto <- data.table::data.table(
  parent = c("a", "b", "b", "b", "c"),
  child = c("b", "c", "d", "e", "f"),
  type = c("part_of", "part_of", "part_of", "is_a", "is_a")
)
weights <- c("part_of" = 0.8, "is_a" = 0.6)
round(calculate_wang_sim_mat(onto, weights = weights), 3)
#>       f     d     b     a     c     e
#> f 1.000 0.470 0.625 0.400 0.796 0.428
#> d 0.470 1.000 0.764 0.477 0.590 0.558
#> b 0.625 0.764 1.000 0.643 0.764 0.742
#> a 0.400 0.477 0.643 1.000 0.477 0.481
#> c 0.796 0.590 0.764 0.477 1.000 0.558
#> e 0.428 0.558 0.742 0.481 0.558 1.000