Machine Learning algorithm Finishing (vii) support vector machine

The stronger the fault tolerance, the better.

B is the plane's biased forward, W is the plane's normal vector, and the X-to-plane mapping:

First of all, the point is the smallest distance from the dividing line, and then ask what kind of W and B, so that the point, the value of the distance dividing line is the largest.

After shrinking:and taking it as min, take yi* (W^t*q (xi) + b) = 1 =

Machine Learning algorithm Finishing (vii) support vector machine