Build a raw co-occurrence sparse matrix — compute_cooc

Constructs a symmetric skill co-occurrence matrix from a data.table of skill assignments. The matrix entry \((i, j)\) counts how many announcements (general_id) mention both skill \(i\) and skill \(j\).

Usage

compute_cooc_matrix(skills)

Arguments

skills: A data.table with columns general_id and escoskill_level_3.

Value

A symmetric sparse matrix of class Matrix::dgCMatrix with skill names as row/column names.

Examples

if (FALSE) { # \dontrun{
skills <- data.table::data.table(
  general_id = c(1, 1, 1, 2, 2),
  escoskill_level_3 = c("A", "B", "C", "A", "B")
)
mat <- compute_cooc_matrix(skills)
} # }