Is convhulln function work for N dimensional

I tried convhulln function on 5 dimentional data. but it shows error in qhullmx "QH6114 qhull precision error: initial simplex is not convex. Distance=-1e+02 The input to qhull appears to be less than 5 dimensional, or a computation has overflowed."

please tell me the limit of N in conhulln fuction. or is their any other function which perform as same as convhulln.

[Pranav Verma]

Hi,

Could you please provide a detailed explanation along with the code you have tried so that we can replicate the issue.

Thanks

Answer 1

[Matt J]

The points you are inputting to convhulln do not form a solid shape in R^5. To within numerical precision, they all lie on a line or a hyper-plane or some sub-dimensional shape like that. As a remedy, you could try this FEX submission,

https://www.mathworks.com/matlabcentral/fileexchange/30892-analyze-n-dimensional-polyhedra-in-terms-of-vertices-or-in-equalities

If P are your input points, you can find the vertices V of their convex hull by doing,

[A,b,Aeq,beq]=vert2lcon(P);

V=qlcon2vert(mean(P),A,b,Aeq,beq)

[He Chris]

Sorry, I'm stiil confued with the usage of these commands.

How can i replace 'convex = convhulln(map_x);' and obtain the variable value 'convex' with the vert2lcon and qlcon2vert ?

Thanks a lot!!

[Matt J]

[A,b,Aeq,beq]=vert2lcon(map_x);

convex=qlcon2vert(mean(map_x),A,b,Aeq,beq)