Quadprog says the problem is non-convex

5 次查看（过去 30 天）
monkey_matlab
monkey_matlab 2016-10-17
编辑： Matt J 2016-10-17
Hello, I am trying to get the solution of the SVM algorithm for the two classes listed. I am trying to use quadprog to solve for the variables alpha1 through alpha6.
Can you tell me what I might be doing wrong with setting up quadprog?
Here is my code:
clear all
close all
clc
Class1 = [1 0; 0.5 1; 0.5 1.5];
Class2 = [0 -1; -1 1; 0.5 -1];
x1_c1 = Class1(:,1);
y1_c1 = Class1(:,2);
x2_c2 = Class2(:,1);
y2_c2 = Class2(:,2);
hold on
axis([-2,2,-2,2])
plot(x1_c1, y1_c1, 'ro');
plot(x2_c2, y2_c2, 'b^');
ClassesA = [Class1; Class2]';
f = [-ones(6,1)]';
b = [zeros(1,6)]';
y = [ones(3,1); -ones(3,1)];
Q = (y * y') .* (ClassesA' * ClassesA);
x = quadprog(Q,y)

请先登录，再回答此问题。

回答（2 个）

Matt J
Matt J 2016-10-17
编辑：Matt J 2016-10-17
Your Q is singular. As one consequence, this makes the problem ill-posed. As another, it makes Q appear numerically to have negative eigenvalues, such that the problem appears non-convex and its minimum unbounded. You must regularize Q in some way, e.g.,
x = quadprog(Q+eye(6)*.000001,y)
  2 个评论
monkey_matlab
monkey_matlab 2016-10-17
编辑：monkey_matlab 2016-10-17
Hello Matt, can you extrapolate your meaning of "regularize Q". Thanks for all your time and help!
Matt J
Matt J 2016-10-17
编辑：Matt J 2016-10-17
By regularize, I mean as discussed here. The example I showed you is a simple example of Tikhonov regularization.

请先登录，再进行评论。

Matt J
Matt J 2016-10-17
If it is natural to put norm constraints on alpha, you can solve with TRUSTREGPROB ( Download ), as opposed to QUADPROG.

类别

在 Help Center 和 File Exchange 中查找有关 Quadratic Programming and Cone Programming 的更多信息

标签

产品

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by