Need help to compute the maximal block of similar objects in a set

Question

Ahmed Hamed 2015-9-16

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/243347-need-help-to-compute-the-maximal-block-of-similar-objects-in-a-set

Given the following table

 Cases       Temperature   Headache    Nausea       Flu
 x1             High        NaN           No        Yes
 x2             Very high   Yes           Yes       Yes
 x3             NaN         No            No        No
 x4             High        Yes           Yes       Yes
 x5             High        NaN           Yes       No
 x6             Normal      Yes           No        No
 x7             Normal      No            Yes       No
 x8             NaN         Yes           NaN       Yes

where xs are set of objects, {Temperature, Headache, Nausea} are set of attributes and {flu} is the only decision attribute I compute the similar object with respect to tolerance relation as the following equation

S_A(x) = {(x,y) \in U x U: \forall a \in A, f_a(x) = f_a(y) or f_a(x) = NaN or f_a(y) = NaN} where f_a(x) is the value of x with respect to attribute a. This leads to (for short i will refer to x1 only)

S_A(x_1) = \{x1,x3,x8} my problem is that i need to split the objects that have different values to find the maximal blocks of similar objects as follows

S_A(x_1) &= {{x1, x3},{x1,x8}}

Here is my code, but i doesn't work very well

 spMBL = zeros();%set of maximal blocks
count = 0;
for i = 1:size(sp,1)
    for j = 1:size(sp, 2)
        if sp(i,j) == 0
            continue;
        end
        ind = sp(i,j);
        for k = 1:size(sp, 2)
            if sp(i,k) == 0
                continue;
            end
                bool = 1;
                for l = 1:size(ind,1)
                    bool = bool * islimited(Data(ind(l),:), Data(sp(i,k),:));
                end
                if bool == 1 && sp(i,j) ~= sp(i,k)
                    ind = [ind;sp(i,k)];
                end
        end
        count = count + 1;
        for m = 1:size(ind,1)
            spMBL(count, m) = ind(m);
        end
        ind = [];
    end
end
for i = 1:size(sp,1)
        if isequal(sp(i,2),0)
            count = count + 1;
            spMBL(count, 1) = sp(i,1);
        end
end
Where Sp is the set of tolerance sets (S_A(x1)...S_A(xn)) Can anyone provide a more powerful pseudocode which could do this (or matlab function)?