Tuesday, October 28, 2008

Hyper-sphere Decision Boundaries


The homework is done for finding a decision boundary with the simplest way between two-class samples.

That is, to find the domain with a center(u) and a radius(r) approaching the decision boundary.











My way to approach the boundary is:
1. Find the centers and the max domains of two training samples individually.
2. Choose the smallest class domain and define the decision boundary.
3. Then the work is done.






























[coding during 10/22~10/24]

No comments: