Neural Network developers : @avo_milas, @ko4osik
ETS decomposition and code aggregation: @rasokurov, @Izy_Golstein
Researcher: @capibaraAttila
- Description: Developed an algorithm to create synthetic time series data that closely resembles real-world trends, intended for use in multiple applications.
- Key Achievements:
- Enhanced the testing and accuracy of forecasting algorithms by expanding the dataset through data augmentation.
- Ensured data confidentiality and anonymity by utilizing synthetic data in sensitive usage scenarios.
- Examples of generated time series:
- Diffusion with ETS decomposition
- Diffusion model only
- Metrics:
