search

ML4ITS / mtad-gat-pytorch

PyTorch implementation of MTAD-GAT (Multivariate Time-Series Anomaly Detection via Graph Attention Networks) by Zhao et. al (2020, https://arxiv.org/abs/2009.02040).
MIT License
328 stars 76 forks source link

some question about model and the result #10

Closed 2snoopy88 closed 3 years ago

2snoopy88 commented 3 years ago

Hi Axel, I have some question for the repo

  1. i read the OmniAnomaly code, i found it use 25/55 as the out_dim for the MSL and SMAP, i also open the MSL data, only the first dimension has value, so can i think the MSL and SMAP is just univariate time series, but almost paper say the dataset is multivariate, it's make me confuse.
  2. i use the repo as the baseline for my research, i found you replace the decoder from VAE to GRU, so did you try the original VAE for the decoder , i do some experiment but i can't achieve the result in the original paper, so if you try VAE can achieve the result in original paper

I will appreciate it if you can reply as soon as possible.

axeloh commented 3 years ago

Hello!

  1. MSL and SMAP has one so-called telemetry value, and then the rest are one-hot encoded values. I don't know if OmniAnomaly tries to forecast/reconstruct the one-hot encoded values, but to us that makes little sense. Therefore, we use all features as input (making it a multivariate), but only forecast/reconstruct the telemetry value.
  2. Yes, we talk about this in #2.
khadijakhaldi commented 1 year ago

Hi @axeloh, if I have 4 columns, each column represent sensor data, and anomaly is based on the 4 features. Can I reconstruct all the 4 columns with your model ? If yes, can you please show me what I need to modify ?