Klassificering – att välja kategori

Drawing the Boundary

~20 min

Classification isn't just about labels—it's about drawing lines (or curves) that separate groups. This line is called a Decision Boundary.

Your Mission

Adjust the line to separate the red points from the blue points. See what happens when the data gets messy ("Overlapping").

Class 0
Class 1
White border = correct, Black border = wrong

Data Mode

Boundary

Metrics

Accuracy:0.0%
Precision:0.0%
Recall:0.0%
TP
0
FP
0
TN
0
FN
0

What Did You Learn?

  • Linear boundaries are limited: A straight line can't separate everything (like the XOR pattern).

  • Real data is messy: When classes overlap, 100% accuracy is impossible.

  • The boundary *is* the model: Training a model just means finding the best place to draw this line.