% Boyd & Vandenberghe "Convex Optimization"
% Joelle Skaf - 10/16/05
% (a figure is generated)
%
% The goal is to find a function f(x) = a'*x - b that classifies the points
% {x_1,...,x_N} and {y_1,...,y_M}. a and b can be obtained by solving a
% feasibility problem:
%           minimize    0
%               s.t.    a'*x_i - b >=  1     for i = 1,...,N
%                       a'*y_i - b <= -1     for i = 1,...,M

% data generation
n = 2;
randn('state',3);
N = 10; M = 6;
Y = [1.5+1*randn(1,M); 2*randn(1,M)];
X = [-1.5+1*randn(1,N); 2*randn(1,N)];
T = [-1 1; 1 1];
Y = T*Y;  X = T*X;

% Solution via CVX
fprintf('Finding a separating hyperplane...');

cvx_begin
    variables a(n) b(1)
    X'*a - b >= 1;
    Y'*a - b <= -1;
cvx_end

fprintf(1,'Done! \n');

% Displaying results
linewidth = 0.5;  % for the squares and circles
t_min = min([X(1,:),Y(1,:)]);
t_max = max([X(1,:),Y(1,:)]);
t = linspace(t_min-1,t_max+1,100);
p = -a(1)*t/a(2) + b/a(2);

graph = plot(X(1,:),X(2,:), 'o', Y(1,:), Y(2,:), 'o');
set(graph(1),'LineWidth',linewidth);
set(graph(2),'LineWidth',linewidth);
set(graph(2),'MarkerFaceColor',[0 0.5 0]);
hold on;
plot(t,p, '-r');
axis equal
title('Simple classification using an affine function');
% print -deps lin-discr.eps
Finding a separating hyperplane... 
Calling sedumi: 16 variables, 3 equality constraints
   For improved efficiency, sedumi is solving the dual problem.
------------------------------------------------------------
SeDuMi 1.21 by AdvOL, 2005-2008 and Jos F. Sturm, 1998-2003.
Alg = 2: xz-corrector, Adaptive Step-Differentiation, theta = 0.250, beta = 0.500
eqs m = 3, order n = 17, dim = 17, blocks = 1
nnz(A) = 48 + 0, nnz(ADA) = 9, nnz(L) = 6
 it :     b*y       gap    delta  rate   t/tP*  t/tD*   feas cg cg  prec
  0 :            9.03E+01 0.000
  1 :   0.00E+00 3.20E+01 0.000 0.3538 0.9000 0.9000  -2.42  1  1  3.8E+01
  2 :   0.00E+00 7.37E+00 0.000 0.2305 0.9000 0.9000  -0.26  1  1  1.4E+01
  3 :   0.00E+00 2.38E-01 0.000 0.0323 0.9900 0.9900   0.62  1  1  5.6E-01
  4 :   0.00E+00 8.82E-06 0.000 0.0000 1.0000 1.0000   0.99  1  1  
iter seconds digits       c*x               b*y
  4      0.0   Inf  0.0000000000e+00  0.0000000000e+00
|Ax-b| =   0.0e+00, [Ay-c]_+ =   0.0E+00, |x|=  0.0e+00, |y|=  5.0e+00

Detailed timing (sec)
   Pre          IPM          Post
1.000E-02    3.000E-02    0.000E+00    
Max-norms: ||b||=0, ||c|| = 1,
Cholesky |add|=0, |skip| = 0, ||L.L|| = 1.
------------------------------------------------------------
Status: Solved
Optimal value (cvx_optval): -0
Done!