ki-ljl / lstm-load-forecasting

Implementation of Electric Load Forecasting Based on LSTM(BiLSTM). Including Univariate-SingleStep forecasting, Multivariate-SingleStep forecasting and Multivariate-MultiStep forecasting.

LSTM-Load-Forecasting

Implementation of Electric Load Forecasting Based on LSTM(BiLSTM). Including Univariate-SingleStep forecasting, Multivariate-SingleStep forecasting and Multivariate-MultiStep forecasting.

Environment

pytorch==1.10.1+cu111

numpy==1.18.5

pandas==1.2.3

Tree

.
│  args.py
│  data_process.py
│  LICENSE
│  models.py
│  README.md
│  tree.txt
│  util.py
│          
├─data
│      data.csv
│      
├─LSTMs
│      multivariate_multi_step.py
│      multivariate_single_step.py
│      univariate_single_step.py
│      
└─models
        multivariate_multi_step.pkl
        multivariate_single_step.pkl
        univariate_single_step.pkl

1. args.py is a parameter configuration file, where you can set model parameters and training parameters.
2. data_process.py is the data processing file. If you need to use your own data, then you can modify the load_data function in data_process.py.
3. Two models are defined in models.py, including LSTM and bidirectional LSTM.
4. util.py defines the training and testing functions of the models in the three prediction methods.
5. The trained model is saved in the models folder, which can be used directly for testing.
6. Data files in csv format are saved under the data file.

Usage

First switch the working path:

cd LSTMs/

Then, execute in sequence:

python multivariate_multi_step.py --epochs 50 --batch_size 30
python multivariate_single_step.py --epochs 30 --batch_size 30
python univariate_single_step.py --epochs 30 --batch_size 30

If you need to change the parameters, please modify them manually in args.py.

Result

lstm-load-forecasting's Issues

about the pred = pred[:, -1, :]

Hi， I am learning your wonderful code.
Now i met a problem about the code : pred = pred[:, -1, :]
Why does just get the last column about the pred?
Could you explain in more detail to me, Thanks a lot!

data source

Hello, if you can, please provide the source of the dataset, thank you very much!

how about multi lstm?

I have the data:

I want use the pre 10 colums to predict the 11th colums,and want use the multi lstm to forecast,as:
model = Sequential()
layers = [1, 75, 100, prediction_steps]
model.add(LSTM(layers[1], input_shape=(None, layers[0]), return_sequences=True)) # add first layer
model.add(Dropout(0.2)) # add dropout for first layer
model.add(LSTM(layers[2], return_sequences=False)) # add second layer
model.add(Dropout(0.2)) # add dropout for second layer
model.add(Dense(layers[3])) # add output layer
model.add(Activation('linear')) # output layer with linear activation
start = time.time()
model.compile(loss="mse", optimizer="rmsprop")
print('Compilation Time : ', time.time() - start)
return model
how shoul i do?

there is no args.py and util.py in your project

Get the future forecast value.

Hello, good work. I have some doubts. The code is a prediction of testdata, and it doesn't output future results. How can I get the values that didn't happen in the future?

where is the args.py, data_process.py, models.py?

可以上传.py文件吗

请问可以将.pyc换成.py文件上传吗？感激不尽

There is .pyc files available there are no .py files available

Hi, can you provide the .py files, because only .pyc files are available and it cannot be opened and also cannot be changed. Thanks

Question for plotting

hi

I have an issue with plotting, why the range of x is from 1 to 151? What is the meaning of the .T[150:300]) in the plot section?

looking forward to your answer
thanks

IndexError: list index out of range

I use my data, have two colums, one is timestamp, one is value.
And use the Multivariate-SingleStep-LSTM and Multivariate-MultiStep-LSTM, the list index out of range: