For benchmark evolution on AFLW2000:
1) In the Table 1 of paper, the reported numbers are evaluated using (x, y) without z ?
1> Evaluation number by your given model hints on that way
2> The methods in that table include 2D methods, which hints on that way
2) For those numbers using (x, y, z) with z, how do you map z from the training domain to the AFLW2000 domain ? when they are in different number scale
cleardusk
commented
5 years ago
For benchmark evolution on AFLW2000: 1) In the Table 1 of paper, the reported numbers are evaluated using (x, y) without z ? 1> Evaluation number by your given model hints on that way 2> The methods in that table include 2D methods, which hints on that way
2) For those numbers using (x, y, z) with z, how do you map z from the training domain to the AFLW2000 domain ? when they are in different number scale