The basic of SVM is NOT basic at all, it is one of the many advanced machine learning algorithms, it belongs to a type of feed forward network for pattern classification and regression problems. It can handle linear as well as non-linear separable patterns. blah...blah... It take serious and formal learning to understand SVM, is one of the many AI modules in the tertiary education. Not for this quick answers forum. Suggest you attend a formal course (I think you are already doing it) and get a good book on neural networks. Meanwhile, enjoy this
video [
^] and be amazed by the power of SVM.