nailo2c/deeplog

In predict.py, I do not understand why "ground_truth" for 'abnormal_has_anomaly' is all 1. Is there no correct or normal data in abnormal data?

Closed this issue · 3 comments

lizhzh8 commented

ground_truth = [1]*len(abnormal_has_anomaly) + [0]*len(normal_has_anomaly) predict = abnormal_has_anomaly + normal_has_anomaly

image

nailo2c commented

Hi @lizhzh8 , that's because I load the abnormal test set first, then load the normal set. And I concat them, so the ground truth set is abnormal on the head and normal on the tail.

lizhzh8 commented

Hi @lizhzh8 , that's because I load the abnormal test set first, then load the normal set. And I concat them, so the ground truth set is abnormal on the head and normal on the tail.

Hi, @nailo2c . In my opinion, there is almost normal data and only a little abnormal data in abnormal dataset.so ground_truth is not all 1 for abnormal_has_anomaly. Is that right?Thanks

nailo2c commented

Hi @lizhzh8 , I think you're right, I should only tag VM in anomaly_labels.txt to be abnormal.
Thanks for your feedback!