What part of the thunderbird dataset did you use ?

LogIntelligence / LogADEmpirical

Log-based Anomaly Detection with Deep Learning: How Far Are We? (ICSE 2022, Technical Track)

MIT License

167 stars 40 forks source link

What part of the thunderbird dataset did you use ? #8

Open SuperCo2 opened 2 years ago

SuperCo2 commented 2 years ago

I am trying to reproduce the results of your paper. In your paper, you wrote"We leverage 10 million continuous log lines..." for the Thunderbird dataset, could you tell me which part you actually used? (I tried using the first 10 million lines but it seemed different)

JinYang88 commented 1 year ago

Same problem here.

xichie commented 1 year ago

I have a similar question. I run the experiment on thunderbird and spirit with 100 logs setting, and the results of logcnn and logrobust seem very poor. In thunderbird, only 816 seqs (1%) are anomalies, the dataset is very imbalanced. How can I get the result in the paper?