I have a sparse matrix:
A=
(14,13) 0.5286
(15,14) 0.6781
(16,15) 0.5683
(17,16) 1.2773
(18,17) 1.0502
(19,18) 0.4966
(21,19) 0.9951
(21,20) 0.4522
(22,21) 0.8507
(23,22) 1.0727
(24,23) 0.8288
(25,24) 0.5811
(26,25) 0.8235
(28,26) 1.5128
(30,28) 0.7966
(30,29) 0.6363
(31,29) 0.8254
(32,31) 0.8573
(33,32) 1.0753
that is result of a minimum spanning tree. now I want to extract 13,14,15,...26,28,29,...33.
as seen 27 is not between numbers. so pred give: 13 14 15 16 17 18 19 21 22 23 24 25 26 28 29 30 31 32 that 20 and 33 is not.
how can I extract total of numbers that say in top?
[ii jj] = find(A);
answer = unique([ii(:); jj(:)]);
should do it.
Note that the find command with two outputs gives you the row and column index of all nonzero elements. Since you have a minimum spanning tree, each number you care about needs to occur at least once in the row or column (for example your matrix never has the number 29 in the first index, but it occurs in the second).
The unique function makes sure that each number that occurs is only represented once.
Related
I have a cell M with n number of cells, with each cell containing several unique numbers, something like this:
{[15 16 21 26 28 145],[2 5 8 9 15],[20 24 27],[10 11 15 8 6 258 74 1],...}
Some of these values appear in more than 1 cell. I would like to calculate the fraction of overlapping values across these cells. For instance, with the 4 cells above, I have 19 unique numbers, and 2 of them belong to more than 1 cell: 15 and 8. Thus, the fraction of overlapping cell is 2/19 = .105. Note that the number of cells in M can vary and thus the number of unique numbers in M also vary as well. Does anyone have any suggestion on how to do this efficiently? I've tried horzcat to concatenate the cells within M then used unique but didn't quite get what I want.
Using the output of the hist() function is useful here.
M = {[15 16 21 26 28 145],[2 5 8 9 15],[20 24 27],[10 11 15 8 6 258 74 1]};
% Bring data to one matrix.
M2 = cell2mat(M);
% Build a histogram from the data with a bin on each unique element. The
% first output of hist is the number of elements in each bin.
a = hist(M2,unique(M2));
% Calculate the overlap by dividing the number of elements that occur more
% than once by the total number of elements.
overlap = sum(a>1)/numel(a);
Hello I'm working with MATLAB and I have a "z" column vector that has dimension of (9680 x 1). I want to reshape it in order to have an array "z" of dimension (44 x 220). I'm doing the following:
z=reshape(z,44,220);
I also tried:
z=reshape(z,[44,220]);
But the output is not right (at least the first row). I can see it by comparing the output matrix with the initial vector.
I just need the 220 first positions of the column vector to be the length of the first row of the matrix, then the next 220 positions of the vector to be the second row of the matrix and so on till obtaining 44 rows.
What am I doing wrong? Thanks for your help.
Matlab stores the matrix values in column major format (this is important during reshape). Since you want row major, you need to do
z = reshape(z, [220 44]).';
i.e. transpose afterwards.
I'd use Andreas H.'s approach.
As an alternative, there's a vec2mat function in the Communications Toolbox that does just that, and even fills missing values if needed:
>> x = 11:18;
>> vec2mat(x,4) %// no padding needed
ans =
11 12 13 14
15 16 17 18
>> vec2mat(x,5) %// padding needed; with 0 by default
ans =
11 12 13 14 15
16 17 18 0 0
>> vec2mat(x,5,-1) %// padding needed; with specified value
ans =
11 12 13 14 15
16 17 18 -1 -1
I have data of integers in x = 500 X 612 matrix. I need a new variable xx in a 500 X 612 matrix but I need to apply cumsum along each row (500) across 12 column steps and applying cumsum like this 51 times --> 500 X (12 X 51) matrix. Then I need a for loop to produce 51 plots of the 500 rows and 12 columns of the cumsum time series. thank you!
I will rephrase what the question is asking to benefit those who are reading.
The OP wishes to segment a matrix into chunks by splitting up the matrix into a bunch of columns. A cumsum is applied to each row individually for each column and are then concatenated together to build a final matrix. As such, given this source matrix:
x =
1 2 3 4 5 6 7 8 9 10 11 12
13 14 15 16 17 18 19 20 21 22 23 24
Supposing that we wish to split up the matrix by columns 3, 6 and 9 and 12, we will have four chunks to work with. We do a cumsum on each of these blocks individually and piece the final result together. So the result would like the following:
xx =
1 3 6 4 9 15 7 15 24 10 21 33
13 27 42 16 33 51 19 39 60 22 45 69
First, you need to determine how many columns you want to break up the matrix into. In your case, we wish to segment the matrix into 4 chunks: Columns 1 - 3, columns 4 - 6, columns 7 - 9, and columns 10 - 12. As such, I'm going to reshape this matrix so that each column is an individual row from a chunk in this matrix. We then apply cumsum over this reshaped matrix and we then reshape it back to what you had originally.
Therefore, do this:
num_chunks = 4; %// Columns 3, 6, 9, 12
divide_point = size(x,2) / num_chunks; %// Determine how many elements are in a row for a cumsum
x_reshape = reshape(x.', divide_point, []); %// Get reshaped matrix
xy = cumsum(x_reshape); %// cumsum over all columns individually
xx = reshape(xy, size(x,2), size(x,1)).'; %// Reconstruct matrix
In the third line of code, x_reshape = reshape(x.', divide_point, []); may seem a bit daunting, but it's actually not that bad. I had to transpose the matrix first because you want to take each row of a chunk and place them into individual columns so we can perform a cumsum on each column. When you reshape something in MATLAB, it collects values column-wise and reshapes the input into an output of a specified size. Therefore, to collect the rows, we need to collect row-wise and so we must transpose this matrix. Next, divide_point tells you how many elements we have for a single row in one chunk. As such, we want to construct a matrix that is of size divide_point x N where divide_point tells you how many elements we have in a row of a chunk and N is the total number of rows over all chunks. Because I don't want to calculate how many there are (am rather lazy actually....), the [] syntax is to automatically infer this number so that we can get a reshaped matrix that respects the total number of elements in the original input. We then perform cumsum on each of these columns, and then we need to reshape this back into the original shape of the input. With this, we use reshape again on the cumsum result, but in order to get it back into the row-order that you want, we have to determine the transpose as reshape takes values in column-major order, then re-transpose that result.
We get:
xx =
1 3 6 4 9 15 7 15 24 10 21 33
13 27 42 16 33 51 19 39 60 22 45 69
In general, the total number of elements to sum over for a row needs to be evenly divisible by the total number of columns that your matrix contains. For example, given the above, if you were to try to segment this matrix into 5 chunks, you would certainly get an error as the number of rows to cumsum over is not symmetric.
As another example, let's say we wanted to break up the matrix into 6 chunks. Therefore, by setting num_chunks = 6, we get:
xx =
1 3 3 7 5 11 7 15 9 19 11 23
13 27 15 31 17 35 19 39 21 43 23 47
You can see that cumsum restarts at every second column, as we desired 6 chunks and to get 6 chunks with a matrix of 12 columns, a chunk is created at every second column.
I have a matrix and i want to consider it has 4 sub matrices which are placed together. How can I find the middle element of each sub matrix when they are together?
consider the matrix below. It is built by 4 sub matrices.
1 2 3 4 5 6
7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 23 24
25 26 27 28 29 30
31 32 33 34 35 36
I want to get their middle elements so i could have:
8, 11, 26, 29
From what I have understood this might work for you and this is a demo, so use your own parameters -
Code
%%// Input matrix
A = rand(44,44);
%%/ Number of submatrices needed
num_submat = 16;%%// 4 for your example case
%%/ Number of submatrices along row and column
num_submat1= sqrt(num_submat);
%%// Middle element indices along each direction
v1 = floor(size(A,2)/(2*num_submat1))+1:size(A,2)/(num_submat1):size(A,2);
%%// Middle elements
middle_ele = A(v1,v1)
It is always helpful to know that matrix indexing in matlab goes columnwise eg,
indOrd = [1,4,7;2,5,8;3,6,9]
where the number is the index order and not related to your example. indOrd(4) would return 4. Try to use this to find the index locations.
Assuming that each submatrix has odd size 2n+1, the coordinates of the center of one submatrix alone are [n+1, n+1]. If your have a square with M*M submatrices (M=2 in your case), the coordinates are [n+1+i*(2*n+1), n+1+j*(2*n+1)], i and j taken independently in the range 0:M-1.
Turning back to Matlab, it is now quite easy to generate all indices of the centers of the submatrices grouped in the matrix A:
n = floor(size(A,1)/(2*M));
xc = n+1+reshape(repmat(0:M-1,M,1),[],1);
yc = n+1+reshape(repmat((0:M-1)',1,M),[],1);
centers = A(yc, xc);
For even-sized submatrix, you have to choose which element is the center, the modification is quite easy to do then.
I have a large number of entries arranged in three columns. Sample of the data is:
A=[1 3 2 3 5 4 1 5 ;
22 25 27 20 22 21 23 27;
17 15 15 17 12 19 11 18]'
I want the first column (hours) to control the entire matrix to create new matrix as follows:
Anew=[1 2 3 4 5 ; 22.5 27 22.5 21 24.5; 14 15 16 19 15]'
Where the 2nd column of Anew is the average value of each corresponding hour for example:
from matrix A:
at hour 1, we have 2 values in 2nd column correspond to hour 1
which are 22 and 23 so the average is 22.5
Also the 3rd column: at hour 1 we have 17 and 11 and the
average is 14 and this continues to the hour 5 I am using Matlab
You can use ACCUMARRAY for this:
Anew = [unique(A(:,1)),...
cell2mat(accumarray(A(:,1),1:size(A,1),[],#(x){mean(A(x,2:3),2)}))]
This uses the first column A(:,1) as indices (x) to pick the values in columns 2 and 3 for averaging (mean(A(x,2:3),1)). The curly brackets and the call to cell2mat allow you to work on both columns at once. Otherwise, you could do each column individually, like this
Anew = [unique(A(:,1)), ...
accumarray(A(:,1),A(:,2),[],#mean), ...
accumarray(A(:,1),A(:,3),[],#mean)]
which may actually be a bit more readable.
EDIT
The above assumes that there's no missing entry for any of the hours. It will result in an error otherwise. Thus, a more robust way to calculate Anew is to allow for missing values. For easy identification of the missing values, we use the fillval input argument to accumarray and set it to NaN.
Anew = [(1:max(A(:,1)))', ...
accumarray(A(:,1),A(:,2),[],#mean,NaN), ...
accumarray(A(:,1),A(:,3),[],#mean,NaN)]
You can use consolidator to do the work for you.
[Afinal(:,1),Afinal(:,2:3)] = consolidator(A(:,1),A(:,2:3),#mean);
Afinal
Afinal =
1 22.5 14
2 27 15
3 22.5 16
4 21 19
5 24.5 15