2. Machine learning Techniques-Dual support Vector machines

Lecture 2. Dual support Vector Machine2.1 motivation of Dual suppor vector machine

The linear support vector machine plus feature transformation will be able to get nonlinear supported vector machine. To do so, we can take advantage of the excellent features of the SVM and feature transformation Q1: The smaller VC Dimension (SVM), the complex boundary (feature Transformation). But this introduces a new problem, the amount of computation is too large 2-1

Figure 2-1

QP has $\tilde{d}$ + 1 variables and N constraints, if the number of variables too much computation is too large.

2.2 Lagrange Dual SVM2.3 solving Dual SVM2.4 Messages behind Dual SVM digression:

2. Machine learning Techniques-Dual support Vector machines