University of Joensuu Pattern Recognition
kNN Classifier

Instructions

You can use in this applet random dataset or make own dataset.

Using random dataset
1. Press button "Random dataset"

Making manual dataset
1. Press the button "Manual dataset creation"
2. Select either Dataset A or dataset B for cluster creation
3. Click cluster points on the white area
4. Press Ready when clusters are complete

Using applet (after random or manual dataset creation)
- Move your mouse over the white area, when you see in real time which points are the nearest.
- Click mouse button on white area when queried point get locked.
- You can change k-value from properties in any time. That means how many nearest points we want to find in the query.
- In results are how many nearest points belongs to dataset A and B. Your point shows that is queried point in dataset A or B.

Note
- Query point's color indicates the dataset which it belongs to.