A tibble containing kanji similarity judgments by 3 "native or native-like"
speakers of Japanese. For each row, the pivot kanji was compared to a list of
potential distractors. From the distractors, the subjects selected one
character which they found particularly easy to confuse with the pivot. For
the exact methodology, see the original study referenced below.
References
Yencken, Lars, & Baldwin, Timothy (2008). Measuring and predicting orthographic associations:
Modelling the similarity of Japanese kanji. In: Proceedings of the 22nd International Conference on Computational
Linguistics (Coling 2008), pp. 1041-1048.