akash13singh / lstm_anomaly_thesis Goto Github PK

Hello Akash. I've some questions for you...

1. Is it possible to run this LSTM code with multi input parameters?
2. The original parameters configuration is the one which have the best results?
3. I don't understand why it returns me 2 very seimilar graphs everytime I run it in PyCharm, could u explain me?
4. On this application it's only possible to run this project by manually labeling the anomalies in the dataset?

Thanks!

Hi! thank you for sharing this amazing code. I have one problem when I try to implement the code.
when I try to run these lines:
multistep_lstm = MultiStepLSTM(look_back=288, look_ahead=1,
layers={'input': 1, 'hidden1':120, 'output':1},
dropout=0.1, loss='mse',
learning_rate=0.02)
model = multistep_lstm.build_model()
it mentioned that: init() missing 1 required positional argument: 'units'

I see you set the self.layers['hidden1'] as units in the function called "StatefulMultiStepLSTM", but set the self.layers['hidden1'] as output_dim in the function"MultiStepLSTM".

can you help me with the missing argument problem?

Hi Akash!
I've read your report and likewise searched within the files and folders of "lstm_anomaly_thesis". the space shuttle dataset was trained and tested but it was not referenced anywhere in the thesis. what's the reason?

whenever i run the lstm_predictor.py I am getting the following error:

FileNotFoundError: [Errno 2] No such file or directory: '/home/sunjeet/lstm_anomaly_thesis-master/logs/run.log'

Its not able to find the log file. Can anyone help me resolve this.

Hi, recently, I am confused by whether the training set should include outliers, I saw you said you trained the normal time series data, I do not know the meaning of normal.

Please reply to me when you saw it,
Many Thanks,
Finn

Hey, I don't know if you would reply to this, but there are files missing like, train_predictions, train_true and so on.
If you could upload them it would help a lot.

Hello Akashm first of all thank you for sharing your work with us, amazing documentation!
I'm trying to run the LSTM code for anomaly detection in my Dataset, but I'm not succeding... Could you give me a little help? My problem is that I dont know how to upload my own data.
I'm a new Python user...

Thanks in advance!