jsyoon0823/TimeGAN
A generative adversarial network for synthesizing realistic time-series data across synthetic and real-world datasets.
Not currently ranked — collecting fresh signals.
star history
TimeGAN implements a framework for generating synthetic time-series data using adversarial training between a generator and discriminator network. The repository includes implementations for multiple dataset types (synthetic sine, stock, and energy data), evaluation metrics including PCA/t-SNE visualization, discriminative scoring via post-hoc classifiers, and predictive metrics. The model architecture supports multiple deep learning modules including GRU, LSTM, and LSTM with layer normalization.
Frequently asked
- What is jsyoon0823/TimeGAN?
- A generative adversarial network for synthesizing realistic time-series data across synthetic and real-world datasets.
- Is TimeGAN open source?
- Yes — jsyoon0823/TimeGAN is an open-source project tracked on heatdrop.
- What language is TimeGAN written in?
- jsyoon0823/TimeGAN is primarily written in Jupyter Notebook.
- How popular is TimeGAN?
- jsyoon0823/TimeGAN has 1.1k stars on GitHub.
- Where can I find TimeGAN?
- jsyoon0823/TimeGAN is on GitHub at https://github.com/jsyoon0823/TimeGAN.