Skip to main content
. Author manuscript; available in PMC: 2020 Aug 10.
Published in final edited form as: IJCAI (U S). 2020 Jul;2020:1395–1402. doi: 10.24963/ijcai.2020/194

Table 1:

Testing accuracies (%) of the original GAT and the GAT applied with each of our 4 CPA models on TRIANGLE-NODE dataset for node classification. We highlight the result of the best performed model. The proportion P of multisets that hold the properties in Theorem 1 among all multisets is also reported.

Dataset
P(%)
TRIANGLE-NODE
29.2
Original 78.40 ± 7.65
Additive 91.31 ± 1.19
Scaled 91.38 ± 1.23
f-Additive 91.18 ± 1.24
f-Scaled 91.36 ± 1.26