Lagrange interpolation polynomial

Question

I have piecewise linear function

I try to build a polynomial of Lagrange for 2, 6, 14 hosts. But for my hosts, my original function become not correct:

I think that my polynomial turns right, but I have troubles to count and build measure of inaccuracy

a = 1;
b = -1;
c = -1;
d = 1;
x = -1:0.33:1;
y = [];
for i = 1 : length(x)
 if (x(i) <= c/2)
    y(i) = x(i)+ 1;
elseif(x(i) >= d/2)
    y(i) = x(i) - 1;
 else
    y(i) = -x(i);
 end
end
plot(x, y);
hold on
nx = -1:0.01:1;
ny=lagrange(x,y,nx);
plot(x,y,'bo');
plot(nx,ny,'r');

function ny=lagrange(x,y,nx)
N = length(x);
ny=zeros(size(nx));
for k=1:N
  t=ones(size(nx));
   for j=[1:k-1, k+1:N]
      t=t.*(nx-x(j))/(x(k)-x(j));
  end
  ny = ny + y(k)*t;
end

Answer 1

Well, I made some mistakes. First of all, I passed on wrong parameters to function. Also, I didn't correctly calculate the values of the original function. The "original" func calculate the values of original graph. The "lagrange" func calculate the values polynomial.

x = -1:0.01:1; 
y = []; 
y = original(x); 
subplot (2, 1, 1); 
plot(x, y); 
xlabel('x'); 
ylabel('y'); 
hold on; 
grid on; 
%hosts
xx = -1:0.5:1;
%values of original func in hosts 
yy = original(xx); 
plot(xx,yy,'o'); 
nx = -1:0.01:1; 
ny=lagrange(xx,yy,nx);     
plot(nx,ny,'r'); 
subplot (2, 1, 2); 
%measure of inaccuracy 
plot(x,ny-y, ''); 
grid on;