I have an "almost symmetric" matrix, which I wish to symmetrize in MATLAB. For example, I wish to symmetrize
>> A = [0 0 1; 2 0 3; 0 3 0]
A =
0 0 1
2 0 3
0 3 0
into
>> B
B =
0 2 1
2 0 3
1 3 0
Safe assumptions are that diagonal entries of A are all zero and that "the bits to change" are always 0. E.g., I changed A(1, 2) and A(3, 1) in the above example, and original values at both locations were 0.
My best attempt based on #Photon's comment (Thanks Photon!) is
>> C = -0.5*(A.'.*A~=0)+1;
>> B = (A+A.').*C
B =
0 2 1
2 0 3
1 3 0
Is there a better (more efficient or faster) way of achieving this?
What about
B = max( A, A.' );
Assuming all entries of A are non-negative.
Related
I have an adjacency matrix M, something like this:
[1 2 0 2 4
2 1 2 0 -1
0 3 1 2 3
2 0 2 1 0
4 -1 3 0 1]
I want to calculate the mean of all values below (but not including) the diagonal. The final output should be 1.5.
To get those values, I thought I'd use N = tril(M,-1). The issue is that I now have zeros in upper and lower part of the matrix N and therefore mean(sum(N)./sum(N~=0)) wouldn't work. Since I also have negative values, I can't just do the mean of values >=0 either. How can I do this?
In one line using logical indexing to extract just the values below the diagonal:
M = [ 1 2 0 2 4;
2 1 2 0 -1;
0 3 1 2 3;
2 0 2 1 0;
4 -1 3 0 1];
mean(M(tril(true(size(M)),-1)))
This returns 1.5 as #excaza indicated.
I have a matrix with some zero values I want to erase.
a=[ 1 2 3 0 0; 1 0 1 3 2; 0 1 2 5 0]
>>a =
1 2 3 0 0
1 0 1 3 2
0 1 2 5 0
However, I want to erase only the ones after the last non-zero value of each line.
This means that I want to retain 1 2 3 from the first line, 1 0 1 3 2 from the second and 0 1 2 5 from the third.
I want to then store the remaining values in a vector. In the case of the example this would result in the vector
b=[1 2 3 1 0 1 3 2 0 1 2 5]
The only way I figured out involves a for loop that I would like to avoid:
b=[];
for ii=1:size(a,1)
l=max(find(a(ii,:)));
b=[b a(ii,1:l)];
end
Is there a way to vectorize this code?
There are many possible ways to do this, here is my approach:
arotate = a' %//rotate the matrix a by 90 degrees
b=flipud(arotate) %//flips the matrix up and down
c= flipud(cumsum(b,1)) %//cumulative sum the matrix rows -and then flip it back.
arotate(c==0)=[]
arotate =
1 2 3 1 0 1 3 2 0 1 2 5
=========================EDIT=====================
just realized cumsum can have direction parameter so this should do:
arotate = a'
b = cumsum(arotate,1,'reverse')
arotate(b==0)=[]
This direction parameter was not available on my 2010b version, but should be there for you if you are using 2013a or above.
Here's an approach using bsxfun's masking capability -
M = size(a,2); %// Save size parameter
at = a.'; %// Transpose input array, to be used for masked extraction
%// Index IDs of last non-zero for each row when looking from right side
[~,idx] = max(fliplr(a~=0),[],2);
%// Create a mask of elements that are to be picked up in a
%// transposed version of the input array using BSXFUN's broadcasting
out = at(bsxfun(#le,(1:M)',M+1-idx'))
Sample run (to showcase mask usage) -
>> a
a =
1 2 3 0 0
1 0 1 3 2
0 1 2 5 0
>> M = size(a,2);
>> at = a.';
>> [~,idx] = max(fliplr(a~=0),[],2);
>> bsxfun(#le,(1:M)',M+1-idx') %// mask to be used on transposed version
ans =
1 1 1
1 1 1
1 1 1
0 1 1
0 1 0
>> at(bsxfun(#le,(1:M)',M+1-idx')).'
ans =
1 2 3 1 0 1 3 2 0 1 2 5
I have a matrix A which contains values 0-100. Matrix B (same number of rows, but only two columns) has values 0 or 1 in its second column. Matrix C has the same number of entries as Matrix A, but just 0 or 1.
I'll assume C starts full of zeros. If there is a 1 in the second column of B, then I want to have a 1 in the same row in C, for every position where there is a value in the corresponding position in A. However, I can't work out how to get this. I've tried using the for loop, but it doesn't seem to be doing what I want.
i.e.
A =
10 10 10
10 10 10
0 20 10
B =
1 0
2 0
3 1
I want to have C:
C =
0 0 0
0 0 0
0 1 1
Very grateful for any help.
Find desired rows using B:
>> B(:, 2) == 1
ans =
0
0
1
Find possible values from A:
>> A ~= 0
ans =
1 1 1
1 1 1
0 1 1
Put it all together by anding using bsxfun:
>> bsxfun(#and, B(:, 2) == 1, A ~= 0)
ans =
0 0 0
0 0 0
0 1 1
Say I have a vector containing only logical values, such as
V = [1 0 1 0 1 1 1 1 0 0]
I would like to write a function in MATLAB which returns a 'streak' vector S for V, where S(i) represents the number of consecutive 1s in V up to but not including V(i). For the example above, the streak vector would be
S = [0 1 0 1 0 1 2 3 4 0]
Given that I have to do this for a very large matrix, I would very much appreciate any solution that is vectorized / efficient.
This should do the trick:
S = zeros(size(V));
for i=2:length(V)
if(V(i-1)==1)
S(i) = 1 + S(i-1);
end
end
The complexity is only O(n), which I guess should be good enough.
For your sample input:
V = [1 0 1 0 1 1 1 1 0 0];
S = zeros(size(V));
for i=2:length(V)
if(V(i-1)==1)
S(i) = 1 + S(i-1);
end
end
display(V);
display(S);
The result would be:
V =
1 0 1 0 1 1 1 1 0 0
S =
0 1 0 1 0 1 2 3 4 0
You could also do it completely vectorized with a couple intermediate steps:
V = [1 0 1 0 1 1 1 1 0 0];
Sall = cumsum(V);
stopidx = find(diff(V)==-1)+1;
V2=V;
V2(stopidx) = -Sall(stopidx)+[0 Sall(stopidx(1:end-1))];
S2 = cumsum(V2);
S = [0 S2(1:end-1)];
Afaik the only thing that can take a while is the find call; you can't use logical indexing everywhere and bypass the find call, because you need the absolute indices.
It's outside the box - but have you considered using text functions? Since strings are just vectors for Matlab it should be easy to use them.
Regexp contains some nice functions for finding repeated values.
Say I have a sparse non-rectangular matrix A:
>> A = round(rand(4,5))
A =
0 1 0 1 1
0 1 0 0 1
0 0 0 0 1
0 1 1 0 0
I would like to obtain the matrix B where the non-zero entries of A are replaced by their linear index in row-first order:
B =
0 2 0 4 5
0 7 0 0 10
0 0 0 0 15
0 17 18 0 0
and the matrix C that where the non-zero entries of A are replaced by the order in which they are found in a row-first search:
C =
0 1 0 2 3
0 4 0 0 5
0 0 0 0 6
0 7 8 0 0
I am looking for vectorized solutions for this problem that scale to large sparse matrices.
If I understand what you are asking, a couple of tranpositions should do the trick. The key is that find(A.') will do "row-first" indexing on A, where .' is the short hand for the transpose of a 2D matrix. So:
>> A = round(rand(4,5))
A =
0 1 0 1 1
0 1 0 0 1
0 0 0 0 1
0 1 1 0 0
then
B=A.';
B(find(B)) = find(B);
B=B.';
gives
B =
0 2 0 4 5
0 7 0 0 10
0 0 0 0 15
0 17 18 0 0
Here's a solution that doesn't require any transposing back and forth:
>> B = A; %# Initialize B
>> C = A; %# Initialize C
>> mask = logical(A); %# Create a logical mask using A
>> [r,c] = find(A); %# Find the row and column indices of non-zero values
>> index = c + (r - 1).*size(A,2); %# Compute the row-first linear index
>> [~,order] = sort(index); %# Compute the row-first order with
>> [~,order] = sort(order); %# two sorts
>> B(mask) = index %# Fill non-zero elements of B
B =
0 2 0 4 5
0 7 0 0 10
0 0 0 0 15
0 17 18 0 0
>> C(mask) = order %# Fill non-zero elements of C
C =
0 1 0 2 3
0 4 0 0 5
0 0 0 0 6
0 7 8 0 0
An outline (Matlab isn't on this machine, so verification is delayed):
You can use find() to get the coordinate list. Let T = A'; [r,c] = find(T)
From the coordinate list, you can create both B and C. Let valB = sub2ind([r,c],T) and valC = 1:length(r)
Use the sparse command to create B and C, e.g. B = sparse(r,c,valB), and then transpose, e.g. B = B' (or could do sparse(c,r,valB)).
Or, as #IanHincks suggests, let B = A'; B(find(B)) = find(B). (I'm not sure why .' is recommended, but, again, I don't have Matlab in front of to check.) For C, simply use C(find(C)) = 1:nnz(A). And transpose back, as he suggests.
Personally, I work with coordinate lists all the time, having migrated away from the sparse matrix representation, just to cut out the costs of index lookups.