search

d-ailin / GDN

Implementation code for the paper "Graph Neural Network-Based Anomaly Detection in Multivariate Time Series" (AAAI 2021)
MIT License
490 stars 142 forks source link

remove some of the 127 features from wadi dataset #94

Closed kaya0007 closed 7 months ago

kaya0007 commented 7 months ago

Hi, thanks for your excellent work! I want to reduce the size of the Wadi dataset. Can I remove any of the 127 features? Is it okay to keep only a part of them? Will this affect the attach label?

looking forward to your reply!

d-ailin commented 7 months ago

Yes, you could remove some of these, but it could potentially lead to performance drop as less information given. To do this, you could just modify the features.txt directly by deleting lines, the program will build the graph based on feature names in the features.txt.

kaya0007 commented 7 months ago

Yes, you could remove some of these, but it could potentially lead to performance drop as less information given. To do this, you could just modify the features.txt directly by deleting lines, the program will build the graph based on feature names in the features.txt.

thanks!