Support Vector Machine

Have spent a lot of time to see the support vector machine, is not started to write, net again slag, I also want to start summing up

　　Support Vector Machine (SVM)

A two-class classification model. The basic model is a linear classifier that defines the maximum interval in the feature space, and the learning strategy is to maximize the interval, and finally the optimization algorithm to solve convex two-times programming.

The linear SVM, linear support vector machine and nonlinear support vector machine are included.

　　Linear scalable support vector machine and maximum hard interval

　　Considering the problem of two kinds of classification, the data is represented by Vector X, and the learning goal of a linear classifier with the category Y ( -1,1) is to find a categorical hyper-plane in the n-dimensional data space, whose equation can be expressed as:

　Ps:w is normal vector, B is intercept

The corresponding classification decision function f (x) is indicated as follows:

F (x) is a linearly-selectable support vector machine , sign (x) is a symbolic function, and x>0 is 1,x<0 to -1,x=0 for 0

　　function interval vs. geometric interval

For a, B, c three points, the farther away from the super-plane, the higher the credibility, thus leading to the concept of function spacing.

function interval:　　　 for a given training set T and the Hyper-plane (w,b), the function interval for defining the Hyper-plane (w,b) about the sample point (Xi,yi) is

Support Vector Machines