Filter löschen
Filter löschen

how to generate random data

11 Ansichten (letzte 30 Tage)
fred bnm
fred bnm am 7 Jan. 2016
Beantwortet: Walter Roberson am 7 Jan. 2016
number of data = 1000; dimension=20; and save in mat format.
  5 Kommentare
jgg
jgg am 7 Jan. 2016
I'm not sure what you mean? What kind of labels do you want assigned?
the cyclist
the cyclist am 7 Jan. 2016
How about this, fred? Spend the time to compose a complete, coherent description of what it is that you want as input and output from your program. This isn't twitter; feel free to use more than 140 characters.

Melden Sie sich an, um zu kommentieren.

Akzeptierte Antwort

Walter Roberson
Walter Roberson am 7 Jan. 2016
data = rand(1000,4);
If the 4 class labels are to be random and equal probability but not necessarily exactly equal number, then
label = randi(4, 1000, 1);
If the 4 class labels must be equal occurrence then
population = [1 * ones(250,1); 2 * ones(250,1); 3 * ones(250,1); 4 * ones(250,1)];
label = population(randperm(1000));
If you wanted a precise distribution that was not equal numbers, then you could adjust the counts in the population matrix.
Your data is 4 dimensional. You cannot plot 4 spatial dimensions. You can encode the 4th dimension as color, but you need to be careful; see http://blogs.mathworks.com/steve/2014/10/20/a-new-colormap-for-matlab-part-2-troubles-with-rainbows/
pointsize = 20;
scatter3(data(:,1), data(:,2), data(:,3), pointsize, data(:,4)); %4th dimension used as color
colormap(parula); %needs R2014b or later
And to add the labels:
labstr = cellstr( num2str(label(:)) );
text( data(:,1), data(:,2), data(:,3), labstr );
this is going to be rather crowded...

Weitere Antworten (0)

Tags

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by