Table 4 Exact matches and matches with phonetic changes in each of the pilot datasets that have been cross-checked for our predictions.

From: Cognate reflex prediction as hypothesis test for a genealogical relation between the Panoan and Takanan language families

Dataset

Intersection

Exact matches

Matches with phonetic changes

Total matches

V &Z

15/44 (34%)

7

6

13/15 (86%)

Oliveira

104/515 (21%)

7

10

17/104 (16%)