Moirai-1.1-R-base
| Property | Value |
|---|---|
| Developer | Salesforce |
| Model Type | Time Series Forecasting |
| Access | Research Only |
| Model URL | https://huggingface.co/Salesforce/moirai-1.1-R-base |
What is moirai-1.1-R-base?
Moirai-1.1-R-base represents a significant evolution in time series forecasting, building upon its predecessor Moirai-1.0-R. This updated version specifically addresses the challenges in predicting low-frequency temporal data, achieving approximately 20% improvement in Normalised Mean Absolute Error (NMAE) for yearly and quarterly datasets from the Monash repository.
Implementation Details
The model demonstrates enhanced capabilities in handling various time series prediction tasks, with particular strength in low-frequency data analysis. It has been validated across 40 datasets from the Monash repository, showing substantial improvements in prediction accuracy.
- Specialized optimization for yearly and quarterly data predictions
- Improved NMAE metrics for low-frequency cases
- Validated across 40 diverse datasets
Core Capabilities
- Enhanced accuracy for low-frequency time series forecasting
- Robust performance on yearly and quarterly data
- Research-focused implementation with academic validation
- Improved normalized mean absolute error metrics
Frequently Asked Questions
Q: What makes this model unique?
The model's distinctive feature is its significant improvement in handling low-frequency time series data, with a 20% enhancement in prediction accuracy for yearly and quarterly datasets compared to its predecessor.
Q: What are the recommended use cases?
The model is specifically designed for research purposes and academic applications, particularly suitable for time series forecasting tasks involving yearly and quarterly data. However, users should carefully evaluate the model's suitability for their specific use case, especially in high-risk scenarios.
