binli826/lstm-autoencoders

Anomaly detection for streaming data using autoencoders

lstm-autoencoders's Issues

sir

I git clone your code, and exec this code in linux Ubuntu use kddcup, there are many bugs, but i fixed some, and get it right. for example ,in you readme, there are also many. And i figure that your code must run in windows envirment, because in code, you forget to change the path "C:/users/libin/xxx" . Can I pull requests to correct this code ?

DataPrepare

Dear Sir.

Hello.

I'm a student studying Machine Learning at Dankook University. While I was studying, I saw your paper called "Anomaly detection in streaming data using autoencoders"(2018).

I'd like to ask for your help, if you don't mind.

If I've seen it right, There was only 'PowerDemand' data not with label. So when I do datapreparation, I don't know how can I apply an option on data.

I am not good at English, so it is awkward to express it, but I would really appreciate it if you could help me.

Thank you.

Yours faithfully.

Request

Hello,
I am so glad to read you master thesis, recently. I'm researching data stream classification in semi-supervised environment and very interested in the work that merged with LSTM.
Can you provide me a data set that could directly running with your public code? very grateful!
my mail is: [email protected].

Wrong FP and FN in parameterHelper.py

When I read the code, found there is an error about FN/FP. "abnomal" is positive, "normal" is negtive, should swap FN with FP.

Online Validation

Hi, very interesting thesis.
I tried to run the code on the forest data set but did not get the same AUC results as you.
I have a doubt about parameters updating in the online part. Is validation step missing? I cannot find the threshold updating in the OnlinePredict. Or is it me doing something wrong?
My email is [email protected]
thank you

help

where can I find datasets published in paper ?