MAD-GANs
MAD-GANs copied to clipboard
LiDan456
The performance of the experiment fall short of the results of the paper
The best of the f1 score that I got is 0.3764 testing with the SWaT data. Is there some hyper parameters that I need to adjust to get close to the score on the paper? (0.77) I trained the model with the default configuration.
Nope
On Sat, Nov 20, 2021 at 9:24 PM Adapter525 @.***> wrote:
Hey, Did you get to implement the code without issues?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/LiDan456/MAD-GANs/issues/6#issuecomment-974744171, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHSUXJRG2ONVWC4YHEYEWBTUNBJ47ANCNFSM4IZHXQTQ . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
Same problem，and each result is different.
@123YuanyuanWang 你找到问题所在了吗？我也AD效果也不好
The best of the f1 score that I got is 0.3764 testing with the SWaT data. Is there some hyper parameters that I need to adjust to get close to the score on the paper? (0.77) I trained the model with the default configuration.
@JuneKyu same. did you figure it out?
结果同样不理想，换其他模型吧。这个论文结果是将阈值不断调整，并测试每一代模型的最好结果。 论文原文也说明了只是做出了时序上的尝试，结果不好。 而且深究一下，代码里面还有不少错误，比如训练与测试集没有用同一个pca模型降维等等。
达利园? @.***
------------------ 原始邮件 ------------------ 发件人: "LiDan456/MAD-GANs" @.>; 发送时间: 2023年3月14日(星期二) 晚上7:39 @.>; @.@.>; 主题: Re: [LiDan456/MAD-GANs] The performance of the experiment fall short of the results of the paper (#6)
Same problem，and each result is different.
@123YuanyuanWang 你找到问题所在了吗？我也AD效果也不好
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>
结果同样不理想，换其他模型吧。这个论文结果是将阈值不断调整，并测试每一代模型的最好结果。 论文原文也说明了只是做出了时序上的尝试，结果不好。 而且深究一下，代码里面还有不少错误，比如训练与测试集没有用同一个pca模型降维等等。 达利园? @.*** … ------------------ 原始邮件 ------------------ 发件人: "LiDan456/MAD-GANs" @.>; 发送时间: 2023年3月14日(星期二) 晚上7:39 @.>; @.@.>; 主题: Re: [LiDan456/MAD-GANs] The performance of the experiment fall short of the results of the paper (#6) Same problem，and each result is different. @123YuanyuanWang 你找到问题所在了吗？我也AD效果也不好 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>
@123YuanyuanWang 我本科毕业论文就是这个题目。用GAN处理时间序列的模型好想不多，TAnoGAN还是处理的单变量时间序列。多变量的时间序列就找到这一个哎。这几天我试了好几次，达不到她论文里面写的结果90几的准确率和召回率。或许代码里哪里可以调整阈值呢？或许还有什么模型比较好呢，愿闻其详