https://github.com/gcarq/keras-timeseries-prediction

Time series prediction with Sequential Model and LSTM units
https://github.com/gcarq/keras-timeseries-prediction

keras lstm machine-learning prediction regression-models

Last synced: 3 months ago
JSON representation

Time series prediction with Sequential Model and LSTM units

• Host: GitHub
• URL: https://github.com/gcarq/keras-timeseries-prediction
• Owner: gcarq
• License: mit
• Created: 2017-01-18T23:25:39.000Z (about 8 years ago)
• Default Branch: master
• Last Pushed: 2017-01-19T00:29:01.000Z (about 8 years ago)
• Last Synced: 2024-10-14T09:48:59.533Z (3 months ago)
• Topics: keras, lstm, machine-learning, prediction, regression-models
• Language: Python
• Homepage:
• Size: 59.6 KB
• Stars: 120
• Watchers: 10
• Forks: 52
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Time series predictions with Keras

#### Requirements

* Theano

* Keras

* matplotlib

* pandas

* scikit-learn

* tqdm

* numpy

#### Usage

```

git clone https://github.com/gcarq/keras-timeseries-prediction.git

cd keras-timeseries-prediction/

pip install -r requirements.txt

python main.py

```

#### Dataset

The dataset is `international-airline-passengers.csv` which contains 144 data points ranging from Jan 1949 to Dec 1960.

Each data point represents monthly passengers in thousands.

![Dataset](data/dataset.png)

#### Model

```

model = Sequential()

model.add(LSTM(64,

               activation='relu',

               batch_input_shape=(batch_size, look_back, 1),

               stateful=True,

               return_sequences=False))

model.add(Dense(1, activation='linear'))

model.compile(loss='mean_squared_error', optimizer='adam')

```

#### Results

![Predictions](data/predictions.png)