Why the reported result of DDGCN in this paper is different with your former paper Orders Are Unwanted: Dynamic Deep Graph Convolutional Network for Personality Detection.

Question

djp0823 opened this issue a year ago · 3 comments

DDGCN result on Kaggle dataset 71.35(71.59) in Paper Orders Are Unwanted: Dynamic Deep Graph Convolutional Network for Personality Detection.

the TrigNet result is also lower than original paper.

Answer 1 · 2024-01-09T03:12:23.000Z

hi, limited by API resources, we randomly selected 200 samples from the test set to evaluate PsyCoT and baselines

Answer 2 · 2024-01-09T07:53:03.000Z

I'm not asking the result of psyCoT, but trained model DDGCN.

Answer 3 · 2024-01-10T04:24:38.000Z

The baselines, including DDGCN and TrigNet, are evaluated on the same randomly selected test samples for fair comparison with PsyCoT.