How to generate pattern randomly In MATLAB

Hello Everyone, I hope you are doing well.

I have the following pattern , i want to generate this pattern randomly in Matlab.

Y axis values is upto 1000 and and X axis values is also 1000 ( thousand samples) ignore the 0.1 value

The current plot has three levels, i want this levels upto 16 which is generated randomly

for example i have three levels then i have three values which are repeated to complete 1000 samples like the plot below.

can anybody please help me

Answer 1

[Davide Masiello]

This is another approach based on the comments under my previous answer.

It tries to implement the request by @Med Future that: it generate the the output y which is equal to number of levelsx1000 for eg. 10 level then shape is 10x1000. but i want the same 10x1000 into single row 1x1000 to generate that shape.

To be fair, I am still not sure this is what @Med Future is trying to achieve, but it might be worth a shot.

NOTE: the code below implements the routine by @Image Analyst to compute random values which have a minimum spacing requirement.

clear,clc

numRequired     = randi(16,1,1);                                            % However many y values you require.
y               = zeros(1, numRequired);
minSpacing      = 30;                                                       % Whatever.
maxIterations   = 100000;                                                   % Way more than you think you'll ever need.  Failsafe to prevent infinite loop.
loopCounter     = 1;                                                        % Number of times the loop goes.  includes keepers and rejects.
counter         = 1;                                                        % Index of y
while counter <= numRequired && loopCounter < maxIterations
	trial_y = 1000 * rand;
	distances = abs(y - trial_y);
	if min(distances) > minSpacing
		% It's far enough away so keep it.
		y(counter) = trial_y;
		counter = counter + 1;
	end
	loopCounter = loopCounter + 1;
end

x = 1:1000;
y = repelem(y,floor(length(x)/length(y)));

% completes y up to 1000 values by repeating the start of y
if length(y) < 1000
    y(end+1:1000) = y(1:1000-end);
end


plot(x,y,'-ob');

[Med Future]

• @Davide Masiello you are doing great job Sir, this method is good, but shape of the pattern is change.
• The pattern shape should be like above which i have shared.
• rest of the method is what i wanted.

[Davide Masiello]

So this?

clear,clc
numRequired     = randi(16,1,1);                                            % However many y values you require.
y               = zeros(1, numRequired);
minSpacing      = 30;                                                       % Whatever.
maxIterations   = 100000;                                                   % Way more than you think you'll ever need.  Failsafe to prevent infinite loop.
loopCounter     = 1;                                                        % Number of times the loop goes.  includes keepers and rejects.
counter         = 1;                                                        % Index of y
while counter <= numRequired && loopCounter < maxIterations
    trial_y = 1000 * rand;
    distances = abs(y - trial_y);
    if min(distances) > minSpacing
        % It's far enough away so keep it.
        y(counter) = trial_y;
        counter = counter + 1;
    end
    loopCounter = loopCounter + 1;
end
x = 1:1000;
y = ones(size(x)).*y';

plot(x,y,'-ob');

Answer 2

[Davide Masiello]

clear,clc

n = 16;                 % Number of levels
L = randi(1000,n,1);    % Level values

x = 1:1000;
y = ones(size(x)).*L;

plot(x,y,'-ob');

[Med Future]

@Davide Masiello Thanks for your answer, but i want random levels upto 16, in your above code we have only 16 level, can you please modified it,

for example in one plot i have two levels , in second plot i have 10 levels. and generate array of 1000 by 1000

[Med Future]

@Davide Masiello and also you can see some random values are closer to one another like one have 965 and other is 958, i want the values to be 30 to 40 % each other for example if i have 965, the other value should be 990

[Image Analyst]

"i want random levels upto 16," <= then make n random.

n = randi(16, 1, 1)   % Number of levels anywhere from 1 to 16

n = 13

[Med Future]

@Image Analyst thanks, The code share above the output y shape depend on number of level for example if i have 10 levels the shape is 10x1000. but i want that 10x1000 in a single row. 1x1000.

and generate a random data which have 1000 rows and each rows have this kind of pattern with different levels.

[Image Analyst]

You can make them separated by a min separation like this. You can generate a trial y value and then get the distance to all the other y values and if it's far enough away, keep it, otherwise try a new y value.

numRequired = 11; % However many y values you require.
y = zeros(1, numRequired);
minSpacing = 30; % Whatever.
maxIterations = 100000; % Way more than you think you'll ever need.  Failsafe to prevent infinite loop.
loopCounter = 1; % Number of times the loop goes.  includes keepers and rejects.
counter = 1; % Index of y
while counter <= numRequired && loopCounter < maxIterations
	trial_y = 1000 * rand;
	distances = abs(y - trial_y);
	if min(distances) > minSpacing
		% It's far enough away so keep it.
		y(counter) = trial_y;
		counter = counter + 1;
	end
	loopCounter = loopCounter + 1;
end
y = sort(y)

y = 1×11

41.3635 140.5341 301.3178 356.8373 425.0306 546.3434 593.6049 632.5164 697.0209 888.9652 966.0485

[Davide Masiello]

The code share above the output y shape depend on number of level for example if i have 10 levels the shape is 10x1000. but i want that 10x1000 in a single row. 1x1000.

and generate a random data which have 1000 rows and each rows have this kind of pattern with different levels.

Could you try to explai this a bit better?

When you say that you want a 10x1000 matrix in a single 1x1000 raw, I am not sure what you mean.

[Rik]

You wanted randomly generated values that were sufficiently spaced, that is the result that Image Analyst provided. This code essentially generates the L variable from the original answer, but calls it y instead.

[Med Future]

@Davide Masiello let me explain it to you.

like the above code share by image Analyst. it generate the the output y which is equal to number of levelsx1000 for eg. 10 level then shape is 10x1000. but i want the same 10x1000 into single row 1x1000 to generate that shape.

secondly i want it random.

[Med Future]

@Rik but the shape has only single value, it does not gives the pattern i want like i share the picture above. which are equally spaced

[Rik]

Look at the x-axis. There are exactly 16 values there. You wanted 16 random levels. The code Image Analyst wrote for you has 16 numbers as a result. What is your problem? You only need to use his code to generate the L matrix in the code Davide wrote for you:

n = 16;                     % Number of levels
L = GenerateLevels(n,30);   % Level values
L=L(:);
x = 1:1000;
y = ones(size(x)).*L;
plot(x,y,'-ob');

function y=GenerateLevels(numRequired,minSpacing)
% numRequired : However many y values you require.
% minSpacing : Whatever.
y = zeros(1, numRequired);
maxIterations = 100000; % Way more than you think you'll ever need.  Failsafe to prevent infinite loop.
loopCounter = 1; % Number of times the loop goes.  includes keepers and rejects.
counter = 1; % Index of y
while counter <= numRequired && loopCounter < maxIterations
	trial_y = 1000 * rand;
	distances = abs(y - trial_y);
	if min(distances) > minSpacing
		% It's far enough away so keep it.
		y(counter) = trial_y;
		counter = counter + 1;
	end
	loopCounter = loopCounter + 1;
end
y = sort(y);
end

See?

Also, you said you wanted random numbers. How can random numbers be equally spaced?

[Med Future]

@Rik like you mention 16 levels, right , i want these 16 levels to be random , like sometime i get 14 sometime i get 4, levels should be random from 2 to 16

[Rik]

Image Analyst also already provided that code for you:

n = randi(16, 1, 1)   % Number of levels anywhere from 1 to 16

You can easily change that range to 2 to 16 like this:

n = 1 + randi(15, 1, 1)   % Number of levels anywhere from 2 to 16

Answer 3

[Steven Lord]

So to clarify you have a vector of y values:

y = [1 2 3];

that you want to repeat over and over to give a pattern like:

yy = [1 2 3 1 2 3 1 2 3 1 2 3] % etc

yy = 1×12

1 2 3 1 2 3 1 2 3 1 2 3

If so use the repmat function.

yy2 = repmat(y, 1, 4)

yy2 = 1×12

1 2 3 1 2 3 1 2 3 1 2 3

If instead you want each element of the resulting vector to contain one of the values from y chosen at random you can use the randi function.

n = numel(y);
ind = randi(n, 1, 12); % A 1-by-12 vector of random integers between 1 and n
yy3 = y(ind) % Use them as indices into y

yy3 = 1×12

3 2 2 2 2 1 2 3 3 3 2 1

Or if you want each element of y to be represented equally often, use randperm to shuffle the elements of yy2.

order = randperm(numel(yy2));
yy4 = yy2(order)

yy4 = 1×12

1 3 2 3 2 1 3 3 2 1 1 2

histogram(yy4) % Show the uniform distribution

If you mean something else, please describe in more detail what data you start with and what your ultimate goal is.

[Med Future]

@Steven Lord The first apporach, yy2 = repmat(y, 1, 4)

• I want random levels 2 to 16,
• then i want random values which is 30 to 40% away from the next value for example one values is 951 the next should be 990.
• Then Repeat the matrix to complete 1000 samples .
• First row consists of one pattern upto 1000 pattern to complete shape of 1000x1000