search

mhkim9714 / MIXAD

MIXAD implementation
MIT License
0 stars 0 forks source link

Questions about the paper (DAEMON model )and the implementation #1

Open yashihuang123 opened 3 months ago

yashihuang123 commented 3 months ago

I've noticed your questions in the discussion forum, and I also have some questions about the paper's implementation. Would you like to exchange ideas with me on this?

mhkim9714 commented 3 months ago

Hi.

Sure we can talk about it. Sorry for the late reply.

2024년 5월 19일 (일) 오후 6:07, yashihuang123 @.***>님이 작성:

I've noticed your questions in the discussion forum, and I also have some questions about the paper's implementation. Would you like to exchange ideas with me on this?

— Reply to this email directly, view it on GitHub https://github.com/mhkim9714/MIXAD/issues/1, or unsubscribe https://github.com/notifications/unsubscribe-auth/AQCXNHVOG3EELWQNTGIZUTDZDBTVLAVCNFSM6AAAAABH6GH5R2VHI2DSMVQWIX3LMV43ASLTON2WKOZSGMYDINJRGE3TGOI . You are receiving this because you are subscribed to this thread.Message ID: @.***>

yashihuang123 commented 3 months ago

Thank you for your response. For the first question: I used this model to test the SMAP data, but found that the F1 score and precision/recall for two files are both 0. In the end, I found that the problem occurred after Fourier transform operations on these two files. Have you ever encountered such a problem? The attached image shows the SMAP results. I look forward to your reply. f1_trends precision_trends For the second question: For the SMD data, the D-rec-loss (Discriminator Reconstruction loss) is very large, and the F-measure is very small and close to 0. I feel that this is not a problem, but I'm not sure. I'd like to consult you about this. I look forward to your reply. M}PA YN$SH{Z8SNS K9HZH0