(Reprinted please indicate the source: http://blog.csdn.net/buptgshengod)
1. We strongly recommend that you read (http://www.cnblogs.com/jerrylead/archive/2011/03/13/1982639.html) support vector Machine SVM (support vector machines) in the background ). SVM is a binary classifier and a popular Classification Algorithm in recent years. This article first introduces some basic knowledge concepts. In the next chapter, we will perform simple code implementation for SVM.
2. Basic Concepts
(1) linear differentiation first introduces what is linear differentiation and references a graph in the previous section. In fact, linear differentiation can be used to separate two different points with a straight line. From this, we can obtain that linear inseparable means that two points are mixed together and cannot be distinguished. However, linear points cannot be separated by mathematical methods. For example, a four-dimensional dataset can be separated by a three-dimensional object called a hyperplane. The blue line.
(2) Support Vector support vector. Now we know the concept of the hyperplane. The support vector is actually the closest vector to the hyperplane. For example, the points closest to the blue line. The method is to determine the distance from the point to the line. Once we find these support vectors, we can zoom in these vectors and only consider these objects, using the idea of minimum sequence optimization.
(3) We need certain constraints for the method of finding Support Vectors Using the Laplace multiplier method. For example, if we set the distance from the vertex to the superplane to d, we need to take the vertex d> 1 as the constraint. Without this constraint, the calculation error may occur. This formula is a set of points with the smallest distance from the point to the superplane.
. When extreme values are obtained under constraints, we use the Laplace multiplier method (see Baidu encyclopedia ).
(4) For the variant, see the formula F (x1, x2,... λ) = f (x1, x2,...)-λ g (x1, x2 ...). Let's change the formula above
The constraint becomes
The above formula uses parameter c to relax the variable, because we can see that some red points in the figure are divided into the green point range. In order to consider this problem, a variable is introduced for control. The main task of svm is to calculate the parameter C.
