HySOM 0.4.0

Fast, lightweight library for training Self-Organizing Maps on 2D time series, tailored for analyzing concentration-discharge hysteresis loops.

HySOM

Fast, lightweight Python library for training Self-Organizing Maps on 2D time series, tailored for analyzing concentration-discharge hysteresis loops.

[!NOTE] Python 3.14 is now supported

Overview

HySOM is a Python library that simplifies the training and visualization of Self-Organizing Maps (SOMs) for 2D time series. It is specifically designed for the study of concentration–discharge (C–Q) hysteresis loops. With HySOM, you can access the General T-Q SOM—a standard framework for classifying sediment transport hysteresis loops. The library also includes several visualization tools to streamline the analysis of sediment transport hysteresis loops. Additionally, HySOM allows you to train your own SOM for C–Q analysis.

---

Features

• Direct access to the General T-Q SOM for sediment transport hysteresis loop analysis
• Tools for analyzing and classifying C-Q hysteresis loops
• Visualization utilities for SOM grids and hysteresis loops
• Easy, yet flexible, training of rectangular Self-Organizing Maps for 2D sequences
• Supports the Dynamic Time Warping distance function
• Lightweight and dependency-minimized

---

The General T-Q SOM

Includes the General T–Q SOM, a standard framework for analyzing sediment transport hysteresis loops. Usage examples can be found in the Documentation

---

📖 Documentation

Comprehensive documentation is provided, including quickstart tutorials, How-to guides and an API reference. Click Here!

---

📦 Dependencies

HySOM requires the following libraries for proper functioning (which are automatically installed when installing HySOM):

• Python
• numpy
• numba
• matplotlib

---

🤝 Contributing

We welcome contributions! If you'd like to include your own standard SOM for C-Q hysteresis analysis, improve the code, report issues, or request features, please open a GitHub issue or pull request.