Interactive K-Nearest Neighbors (KNN) Sandbox

Class Red

Class Blue

Query Point (X)

Distance Radius

Confusion Matrix (Leave-One-Out Evaluation)

Predicted Red (+)

Predicted Blue (-)

Total

Actual Red

0 0.0% True Red (TP)

0 0.0% False Blue (FN)

Actual Blue

0 0.0% False Red (FP)

0 0.0% True Blue (TN)

Total

1. Left-Click Canvas Action

+ Add Red + Add Blue

🎯 Move Query Point

2. Number of Neighbors (K)

K Neighbors Value3

Tip: Use odd numbers for binary classification to avoid tie votes!

Ensemble Vote Tally Matrix
Closest Neighbors found: 0
Vote Breakdowns: N/A

🔬 Classroom Demonstration: Put the app in "Move Query Point" mode. Place it directly inside a dense blue cluster. Now, ramp K up to **15**. Notice how the point can switch to Red if there are more red points globally nearby, showcasing how a large K underfits local structures! The system Confusion Matrix updates automatically based on a global Leave-One-Out validation strategy.