Precision Score

Measures how many retrieved items are relevant (i.e. in ground truth).

• Inputs: retrieved list and ground truth list
• Returns: precision as a float between 0 and 1

Example

from vero.metrics import PrecisionScore

#example inputs
#ch_r is the retrieved citations from the retriever
#ch_t is the ground truth citations
ch_r = [1,2,3,5,6]
ch_t = [2,3,4]
ps = PrecisionScore(ch_r, ch_t)
print(ps.evaluate())

Output

0.60