Functime

What is Functime?

Functime is an automated machine learning (AutoML) tool tailored for business forecasting, characterized by its high speed and efficiency in processing time series data. It stands out for its rapid execution, capable of forecasting, backtesting, and scoring 100,000 time series in under 10 seconds. Functime integrates various external data sources, including holiday, weather, and economic data, facilitating a comprehensive analysis beyond numerical trends. The tool offers advanced techniques such as quantile regression, conformal prediction, and GPT-4-enabled trend-seasonality analysis for forecasting. Additionally, Functime provides feature importance plots, Shapley values, and sensitivity analysis for thorough forecast interpretation. Visualizations generated through Functime using ECharts and Plotly libraries can be easily shared and embedded into products for streamlined communication and analysis.

Who created Functime?

Functime was created by TracecatHQ. It was launched on June 18, 2024. Functime is an automated machine learning (AutoML) tool designed specifically for business forecasting. It is capable of forecasting, backtesting, and scoring 100,000 time series in under 10 seconds. The tool incorporates various data sources and advanced techniques such as quantile regression and conformal prediction for forecasting. Additionally, Functime offers GPT-4-enabled trend-seasonality analysis and provides powerful visualization capabilities using ECharts and Plotly libraries.

What is Functime used for?

• Fast feature engineering
• Supports embarrassingly-parallel forecasts
• Robust probabilistic forecasting
• Quantile regression techniques
• Conformal prediction techniques
• Handles zero-inflated datasets
• Handles censored forecasts
• Blazing-fast hyperparameter tuning
• Scores forecasts in parallel
• Supports numerous metrics

Who is Functime for?

• Forecast Analysts
• Data scientists
• Business analysts
• Machine learning engineers
• Business stakeholders
• Data Analysts

How to use Functime?

To use Functime, follow these steps:

1. Install Functime by running pip install functime.
2. Explore the GitHub repository and documentation provided for a comprehensive guide on using Functime.
3. Start with the tutorial to get familiar with the tool and create your first forecasting pipeline.
4. Utilize the evaluation procedures for scoring, ranking, and plotting functions to assess forecasts efficiently.
5. Leverage the LLM Forecast Analysts feature to analyze and compare trends, seasonality, and causal factors in forecasts.
6. Refer to the API reference for detailed information on Functime's API for more advanced usage.
7. Integrate external data such as holiday, weather, and economic data to enhance forecasting accuracy.
8. Use fast feature engineering capabilities for quick data processing and analysis.
9. Benefit from robust probabilistic forecasting, quantile regression techniques, and specialized options for zero-inflated and censored datasets.
10. Explore the GPT-4-enabled trend-seasonality analysis for detailed prediction justifications.
11. Generate visualizations using ECharts and Plotly libraries to represent forecasts and analysis results.
12. Easily share visualizations with others, including direct embedding into products for seamless communication.

By following these steps, users can effectively utilize Functime for fast and accurate business forecasting with advanced features and visualization capabilities.