May I ask why the test results of last.fm data set are about 10% higher than those in the paper about kgcn-pytorch HOT 17 CLOSED

i dont know

from kgcn-pytorch.

You can share the metrics code that you wrote？thank you very much。

from kgcn-pytorch.

可以分享一下你写的实验结果评价标准代码？非常感谢。

from kgcn-pytorch.

可以分享一下你写的实验结果评价标准代码？非常感谢。

I used his code without any modifications.
论文里面的评价指标是AUC和F1，他的代码实验结果评价指标只有 损失值和roc_auc_score，是怎么看出来比论文里面的结果好10%呀？

from kgcn-pytorch.

May I ask why the test results of last.fm data set are about 10% higher than those in the paper

可以分享一下你写的实验结果评价标准代码？非常感谢。

from kgcn-pytorch.

我也没改，按照源代码跑的结果，具体明天我看一下

from kgcn-pytorch.

好的，感谢感谢

from kgcn-pytorch.

from kgcn-pytorch.

roc_auc_score算出来的结果就是Auc的值，我跑出来的结果roc_auc_score=0.89568，F1=0.80918
…
---Original--- From: @.> Date: Tue, Aug 2, 2022 22:19 PM To: @.>; Cc: @.>;"State @.>; Subject: Re: [zzaebok/KGCN-pytorch] May I ask why the test results of last.fmdata set are about 10% higher than those in the paper (Issue #5) 好的，感谢感谢 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: @.***>

嗯嗯，感谢感谢，可以分享一下F1是怎么写的吗

from kgcn-pytorch.

from kgcn-pytorch.

from sklearn.metrics import roc_auc_score,f1_score这两个函数，你跑出来结果怎么样呢，可以告诉我一下吗
…
---Original--- From: @.> Date: Wed, Aug 3, 2022 16:58 PM To: @.>; Cc: @.>;"State @.>; Subject: Re: [zzaebok/KGCN-pytorch] May I ask why the test results of last.fmdata set are about 10% higher than those in the paper (Issue #5) roc_auc_score算出来的结果就是Auc的值，我跑出来的结果roc_auc_score=0.89568，F1=0.80918 … ---Original--- From: @.> Date: Tue, Aug 2, 2022 22:19 PM To: @.>; Cc: @.>;"State @.>; Subject: Re: [zzaebok/KGCN-pytorch] May I ask why the test results of last.fmdata set are about 10% higher than those in the paper (Issue #5) 好的，感谢感谢 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: @.> 嗯嗯，感谢感谢，可以分享一下F1是怎么写的吗 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: @.>

Auc好像是0.86左右，f1没有跑，每天加上f1再跑一边，给你具体结果

from kgcn-pytorch.

Auc是0.85610，f1是0.779465

from kgcn-pytorch.

from kgcn-pytorch.

我也不知道

from kgcn-pytorch.

请问您能分享一下加入了AUC，F1score的文件或者代码吗？我加入了评价指标的代码，但是跑出来超级低，我不知道哪里出了问题。能否看一下您的代码？非常感谢

from kgcn-pytorch.

请问您能分享一下加入了AUC，F1score的文件或者代码吗？我加入了评价指标的代码，但是跑出来超级低，我不知道哪里出了问题。能否看一下您的代码？非常感谢

暂时不用了，我知道了，我按照论文设置的batchsize跑出来很低auc0.5，按照本复现代码的作者设置的batchsize跑出来很好。

from kgcn-pytorch.

outputs_predict = [int(item > 0.5) for item in outputs.cpu().detach().numpy()]
total_f1 += f1_score(labels.cpu().detach().numpy(), outputs_predict)
print('[Epoch {}]test_f1: '.format(epoch + 1), total_f1 / len(test_loader))
这是我加的f1

from kgcn-pytorch.