Data field shifting through a vector of data in matlab - arrays

I need to create a data field that will go through a vector. Data field is constant length, and it is going through the data vector shifting data field with data field length. I need the mean value of that field (A vector) that corresponds to a mean value of another field (B vector).
Example:
A=[1 5 7 8 9 10 11 13 15 18 19 25 28 30 35 40 45 48 50 51];
B=[2 4 8 9 12 15 16 18 19 20 25 27 30 35 39 40 45 48 50 55];
I want to do next:
A=[{1 5 7 8 9} 10 11 13 15 18 19 25 28 30 35 40 45 48 50 51];
B=[{2 4 8 9 12} 15 16 18 19 20 25 27 30 35 39 40 45 48 50 55];
I want to take data from field of 5 points and get mean value. And then shift whole data field with data field length.
A=[1 5 7 8 9 {10 11 13 15 18} 19 25 28 30 35 40 45 48 50 51];
B=[2 4 8 9 12 {15 16 18 19 20} 25 27 30 35 39 40 45 48 50 55];
I need two vectors, C and D with mean values of this method.
C=[6 13.4 27.4 45.2];
D=[7 17.6 31.2 47.6];
I started something with
n = length(A);
for k = 1:n
....
but nothing I tried worked.


reshape the vector into a 5-row matrix and then compute the mean of each column:
C = mean(reshape(A,5,[]),1);
D = mean(reshape(B,5,[]),1)

Related

How do I perform shell sort using sequence {3,2,1}?

Suppose I have an array:
30 20 29 19 28 18 27 17 26 16 25 15 24 14 23 13 22 12 21 11
I am not understanding how to do the shell sort using sequence 3:
Would I simply do this:
30 20 29 19 28 18 27 17 | 26 16 25 15 24 14 |23 13 22 12 21 11
Where I split it into 3 parts and sort the respective parts? And then do the same with 2 sorting after, except split into halves? What is the right way to do this? Can someone please explain?

If you look at your array and number the locations
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
30 20 29 19 28 18 27 17 26 16 25 15 24 14 23 13 22 12 21 11
In a shell sort, what you do is start with a skip number (in your case 3) so to make the first "list" you take a number and skip. With 3 this would be 1st, 4th, 7th etc.
So you would have a list of
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
30 19 27 16 24 13 21
and a second list of
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
20 28 17 25 14 22 11
The 3rd list is the remaining items.
For the next round you do it with one less... so items at odd number locations and items at even number locations.
In response to comment below
A Shell sort is an in-place sort — that means you don't remove the items to new lists or create any new data structures. You are using the array to "treat" items which are "far apart" in terms of array locations as next to each other. You don't actually make new lists or new arrays (that is why I showed my diagrams as I did); you just look at these locations.
Why?
Because it means that when you start (for example with 3) you are moving stuff farther) -- eg the 13 that starts at location 16 gets moved to location 1 in the first pass. Then as you reduce the number you start doing more local changes. This means you gain an advantage over a typical bubble sort. Still not good — but MUCH better than a bubble sort.

Create matrix of some repeating numbers

EN = 10;
etable = [1,2,3,4,5,6;
4,5,6,7,8,9;
7,8,9,10,11,12;
10,11,12,13,14,15;
13,14,15,16,17,18;
16,17,18,19,20,21;
19,20,21,22,23,24;
22,23,24,25,26,27;
25,26,27,28,29,30;
28,29,30,31,32,33];
Is it possible to make a for loop in which I just change the EN value and it automatically creates etable? Because I have to make EN 50 so it'll not be good to write 50 lines in etable.
I tried in this way:
EN = 10;
c = 1:EN;
nodes = zeros(size(c',1),2);
for i = 1:length(c)
nodes(i,1) = i;
nodes(i,2) = i+1;
end
etable = zeros(size(c',1),6);
for i = 1:size(nodes,1)
etable(i,1) = 2*nodes(i,1)-1;
etable(i,2) = 2*nodes(i,1);
etable(i,3) = 2*nodes(i,1)+1;
etable(i,4) = 2*nodes(i,2);
etable(i,5) = 2*nodes(i,2)+1;
etable(i,6) = 2*nodes(i,2)+2;
end

You can use implicit expansion implicitly (introduced in MATLAB R2016b) or explicitly by using MATLAB's bsxfun to create that matrix:
% Parameters
EN = 10;
n = 6;
step = 3;
% Matrix (implict expansion, MATLAB >= R2016b)
etable = (0:step:EN*step-1).' + (1:n)
% Matrix (bsxfun, MATLAB < R2016b)
etable_legacy = bsxfun(#plus, (0:step:EN*step-1).', (1:n))
For the given parameter set, the outputs are:
etable =
1 2 3 4 5 6
4 5 6 7 8 9
7 8 9 10 11 12
10 11 12 13 14 15
13 14 15 16 17 18
16 17 18 19 20 21
19 20 21 22 23 24
22 23 24 25 26 27
25 26 27 28 29 30
28 29 30 31 32 33
etable_legacy =
1 2 3 4 5 6
4 5 6 7 8 9
7 8 9 10 11 12
10 11 12 13 14 15
13 14 15 16 17 18
16 17 18 19 20 21
19 20 21 22 23 24
22 23 24 25 26 27
25 26 27 28 29 30
28 29 30 31 32 33
You can change EN (number of rows), the "number of columns" n and/or the "step between rows" step.
Hope that helps.

2D matrix to 3D matrix with row to [row, col] mapping

I have a 2D matrix with in the 1st dimension different channels, and in the 2nd dimension time samples. I want to rearrange this to a 3D matrix, with in the 1st and 2nd dimension channels, and in the 3rd time samples.
The channels have to mapped according to a certain mapping. Right now I am using a for-loop to do so, but what would be a no-loop solution?
N_samples = 1000;
N_channels = 64;
channel_mapping = reshape(1:64, [8 8]).';
% Results in mapping: (can also be random)
% 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 37 38 39 40
% 41 42 43 44 45 46 47 48
% 49 50 51 52 53 55 55 56
% 57 58 59 60 61 62 63 64
data = rand(N_channels, N_samples);
data_grid = NaN(8,8, N_samples);
for k = 1:N_samples
tmp = data(:, k);
data_grid(:, :, k) = tmp(channel_mapping);
end

You can do it in one go as follows:
data_grid = reshape(data(channel_mapping, :), 8, 8, []);

Find first and last of a unique element in a column

In a data table such as with the following format:
id Time 1 Time2 V1 V2
1 1 10 30 40
1 2 20 31 41
1 3 30 32 42
1 4 40 33 43
2 1 10 40 50
2 2 20 41 51
2 3 30 42 52
2 4 40 43 53
3 1 10 50 60
3 2 20 51 61
3 3 30 52 62
3 4 40 53 63
I want to select the two smallest and two largest variable readings of time 1 and time 2
I want to do a regression and correlation analysis of v1 and v2 using the first two and last two time readings for each unique ID
Thanks

Alternative shorter method to permute and reshape a multidimensional-array

I have the following multidimensional array:
a(:,:,1) =
11 12 13
21 22 23
a(:,:,2) =
21 22 23
31 32 33
a(:,:,3) =
31 32 33
41 42 43
The end result looks as follows:
e =
11 21 31
12 22 32
13 23 33
21 31 41
22 32 42
23 33 43
In order to obtain array e, I did the following:
b=permute(a,[2,1,3])
c=reshape(b,1,6,[])
d=permute(c,[2,1,3])
Is there a cleaner or more elegant way of achieving the same result?

Here permute used once:
b = permute(a,[2 1 3]);
result = reshape(b, [], 3)

You can do it like the following:
reshape([a(:,:,1);a(:,:,2);a(:,:,3)].',6,3)

Resources