Cookbook

Short, task-focused recipes you can copy and run. Each recipe is self-contained and uses small, seeded datasets so it runs in seconds. The code blocks are executed in CI to keep them working; the text blocks below them show illustrative output and may vary slightly across platforms and dependency versions. For conceptual background see the User Guides; for the full API see the API Reference.

• Quickstart: train, predict, and score — the basic workflow plus scikit-learn Pipeline, cross_val_score, and GridSearchCV.
• Choosing a model — the TSK families and how to swap between them.
• Feature selection on high-dimensional data — embedded feature selection and rule extraction with DGTSKClassifier.
• Persistence and the MF cache — save/load and managing the membership-function cache.
• Inspecting a trained model — rules, membership functions, activations, and feature importance.