
Catalyst
PyTorch framework for Deep Learning R&D focusing on reproducibility and rapid experimentation.

Lightning fast forecasting with statistical and econometric models.

StatsForecast is a Python library designed for high-performance time series forecasting. It leverages statistical and econometric models, including ARIMA, ETS, CES, and Theta, optimized with numba for speed. The architecture is built for scalability, enabling efficient fitting of millions of time series. It supports integration with Spark, Dask, and Ray for distributed computing. StatsForecast provides probabilistic forecasting, confidence intervals, exogenous variables, and anomaly detection. It is compatible with sklearn syntax, making it easy to use. Key value propositions include speed, accuracy, and scalability, making it suitable for production environments and benchmarking. It allows users to forecast in production environments or as benchmarks. The library offers an extensive set of models that efficiently fit millions of time series.
StatsForecast is a Python library designed for high-performance time series forecasting.
Explore all tools that specialize in model evaluation. This domain focus ensures StatsForecast delivers optimized results for this specific requirement.
Automatically searches for the best parameters and selects the optimal model (AutoARIMA, AutoETS, etc.) for time series data.
Provides confidence intervals and prediction intervals for forecasts, quantifying uncertainty.
Allows the inclusion of external factors (e.g., weather, prices) to improve forecast accuracy.
Detects unusual patterns in time series data using in-sample prediction intervals.
Handles time series data with multiple seasonal patterns using MSTL (Multiple Seasonality Time series decomposition using Loess).
Install StatsForecast: `pip install statsforecast` or `conda install -c conda-forge statsforecast`
Import necessary modules: `from statsforecast import StatsForecast`
Load your time series data into a DataFrame.
Define your forecasting models: `from statsforecast.models import AutoARIMA`
Initialize StatsForecast with your models and frequency: `sf = StatsForecast(models=[AutoARIMA(season_length=12)], freq='ME')`
Fit the models to your data: `sf.fit(df)`
Make predictions: `sf.predict(h=12, level=[95])`
All Set
Ready to go
Verified feedback from other users.
"StatsForecast is praised for its speed, accuracy, and scalability in time series forecasting."
Post questions, share tips, and help other users.

PyTorch framework for Deep Learning R&D focusing on reproducibility and rapid experimentation.

A unified platform for building, deploying, and managing AI agent systems securely.
Supervise.ly provides an all-in-one platform for computer vision, enabling users to curate, label, train, evaluate, and deploy models for images, videos, 3D, and medical data.