Non-Parametric Modelling

Let the data speak for itself.

Introduction

Non-parametric models make fewer assumptions about the underlying data distribution, making them powerful for complex real-world patterns.

What You Will Learn

• Implement k-Nearest Neighbours for classification and regression
• Apply kernel methods and SVM
• Master advanced tree-based methods
• Understand when non-parametric methods outperform parametric ones

Assessment Connection

Section A — non-parametric methods (especially tree-based) are often the strongest choice for tabular workplace data.

📘 Tutorials

Step-by-step guides to implementing models:

• k-Nearest Neighbours (k-NN)
• Naive Bayes
• SVM and Kernel Methods
• Advanced Tree-Based Methods
• XGBoost & LightGBM

---

🛠️ How-To Guides

Practical, goal-oriented instructions:

• How to Choose the Right \(k\) Value
• How to Select an SVM Kernel
• How to Prune and Regularise Decision Trees

---

📖 Reference

Quick lookups and technical specifications:

• Non-Parametric Comparison
• XGBoost & LightGBM Parameters

---

🧠 Explanation

Deep dives into fundamental concepts:

• Parametric vs. Non-Parametric Models
• The No Free Lunch Theorem