jaccard is very useful to gauge the amount of overlap. this pull request adds n_intersections to the outputs that shows the number of unique intersections in -a that overlap with -b.
between the bases of overlap (intersection)
the proportion of overlap (jaccard)
and now the number of intersections (n_intersections)
this gives the most commonly used (simple) metrics when comparing 2 datasets.
arq5x
commented
11 years ago
jaccard is very useful to gauge the amount of overlap. this pull request adds n_intersections to the outputs that shows the number of unique intersections in -a that overlap with -b.
between the bases of overlap (intersection)
the proportion of overlap (jaccard)
and now the number of intersections (n_intersections)
this gives the most commonly used (simple) metrics when comparing 2 datasets.