pyfuzzy-toolbox

A comprehensive Python library for Fuzzy Systems with focus on education and professional applications. Includes inference, learning, fuzzy differential equations, and p-fuzzy systems.

📚 Documentation

Read the full documentation →

• Getting Started: Installation and quick tutorials
• User Guides: In-depth guides for each module
• API Reference: Complete method documentation
• Examples: 18+ interactive Colab notebooks

📦 Installation

Basic Installation (Library only)

pip install pyfuzzy-toolbox

Full Installation (with Web Interface)

For the complete experience including the interactive web interface:

macOS/Linux (Zsh/Bash):

pip install 'pyfuzzy-toolbox[ui]'

Windows (PowerShell/CMD):

pip install pyfuzzy-toolbox[ui]

Note: Package name is pyfuzzy-toolbox, import as fuzzy_systems:

import fuzzy_systems as fs

🖥️ Web Interface (NEW!)

Launch the interactive web interface with a single command:

pyfuzzy interface

This opens a modern Streamlit-based interface featuring:

• 🎯 ANFIS: Complete workflow (data → training → evaluation → prediction)
• 📊 Interactive visualizations: Membership functions, decision surfaces, training curves
• 🔮 Real-time predictions: Manual input or batch CSV upload
• 📈 Model analysis: Rules visualization, feature importance, sensitivity analysis
• 💾 Export capabilities: Models, predictions, and results

CLI Commands

# Launch web interface (browser opens automatically)
pyfuzzy interface

# Custom port and dark theme
pyfuzzy interface --port 8080 --dark-theme

# Headless mode (no browser)
pyfuzzy interface --no-browser

# Show version
pyfuzzy version

# Help
pyfuzzy --help

Programmatic API

Launch the interface from Python code or Jupyter notebooks:

from fuzzy_systems import launch_interface

# Simple launch
launch_interface()

# Custom configuration
launch_interface(
    port=8080,
    theme='dark',
    open_browser=True
)

🧩 Core Modules

fuzzy_systems.core

Fundamental fuzzy logic components

• Membership functions: triangular, trapezoidal, gaussian, sigmoid, generalized_bell
• Classes: FuzzySet, LinguisticVariable
• Operators: fuzzy_and_min, fuzzy_or_max, fuzzy_not

fuzzy_systems.inference

Fuzzy inference systems

• MamdaniSystem: Classic fuzzy inference with defuzzification (COG, MOM, etc.)
• SugenoSystem: TSK systems with functional outputs (order 0 and 1)

fuzzy_systems.learning

Learning and optimization

• ANFIS: Adaptive Neuro-Fuzzy Inference System
• WangMendel: Automatic rule generation from data
• MamdaniLearning: Gradient descent and metaheuristics (PSO, DE, GA)

fuzzy_systems.dynamics

Fuzzy dynamic systems

• FuzzyODE: Solve ODEs with fuzzy uncertainty (α-level method)
• PFuzzySystem: Discrete and continuous p-fuzzy systems

📓 Interactive Notebooks

Explore hands-on examples organized by topic:

All notebooks can be opened directly in Google Colab!

📖 Quick Start Guides

Comprehensive guides for each module with theory, examples, and best practices:

🎛️ Inference Systems

Build fuzzy control systems and decision-making tools.

🧠 Learning & Optimization

Automatic rule generation and parameter tuning from data.

🌊 Dynamic Systems

Model temporal evolution with fuzzy uncertainty.

📚 View All Guides

⚡ Quick Example

import fuzzy_systems as fs

# Create Mamdani system
system = fs.MamdaniSystem()
system.add_input('temperature', (0, 40))
system.add_output('fan_speed', (0, 100))

# Add terms
system.add_term('temperature', 'cold', 'triangular', (0, 0, 20))
system.add_term('temperature', 'hot', 'triangular', (20, 40, 40))
system.add_term('fan_speed', 'slow', 'triangular', (0, 0, 50))
system.add_term('fan_speed', 'fast', 'triangular', (50, 100, 100))

# Add rules
system.add_rules([('cold', 'slow'), ('hot', 'fast')])

# Evaluate
result = system.evaluate(temperature=25)
print(f"Fan speed: {result['fan_speed']:.1f}%")

🔗 Links

• Documentation: https://1moi6.github.io/pyfuzzy-toolbox/
• PyPI: https://pypi.org/project/pyfuzzy-toolbox/
• GitHub: https://github.com/1moi6/pyfuzzy-toolbox

📝 Citation

@software{pyfuzzy_toolbox,
  title = {pyfuzzy-toolbox: A Comprehensive Python Library for Fuzzy Systems},
  author = {Cecconello, Moiseis},
  year = {2025},
  url = {https://github.com/1moi6/pyfuzzy-toolbox},
  note = {Includes inference, learning, fuzzy differential equations, and p-fuzzy systems}
}

📄 License

MIT License - see LICENSE for details.