Why is the F1 score run of your code so much higher than what the original paper predicted? The jazz dataset in the paper only has an F1 score of 0.08 under the SIR model, so why does your code run GCNSI with an F1 score of 0.8+? The rest of the dataset also performs much better than the paper's data, is there any optimization for the paper's model? Looking forward to receiving a reply.
Why is the F1 score run of your code so much higher than what the original paper predicted? The jazz dataset in the paper only has an F1 score of 0.08 under the SIR model, so why does your code run GCNSI with an F1 score of 0.8+? The rest of the dataset also performs much better than the paper's data, is there any optimization for the paper's model? Looking forward to receiving a reply.