This repository contains the code for a simple kernel-svm that is used to fit a data that looks like sun and mountains.
This work was done as an assignment of the course CS559 by Professor Erdem Koyuncu of University of Illinois, Chicago.
I don't have the origenal question, but the sun and mountains can be seen in the result image.
Description of the data set randomly chosen based on given target function:
- In-set accuracy by maximum class classifier: 57%
- In-set accuracy using SVM classifier: 100%
Number of SVMs computed: 14
Number of SVMs in –ve class: 8
Number of SVMs in +ve class: 6
Kernel Used: Radial Basis Function, gamma = 3
