different outputs for same input in every run

Dear all,

I am using fitcknn function and I expect a 8*2 matrix to be the output (detected varialble). But sometime the output is correct and other times not. I used fixed rand number seed and clear all variables before every run but it does not work.

Any idea what is the problem of my code? Many thanks in advance.

clear all;
data = load('data.mat').data;


%%%%%%%%%
find_kmeans=kmeans(data(:,1),2);
data_2=table(data(:,1),find_kmeans(:));
data_2.Properties.VariableNames = ["counts", "detect"];

Mdl=fitcknn(data_2,'detect');
Mdl.NumNeighbors=1;

test_data=data(:,1);
pred=predict(Mdl,test_data);

wanted_points=(find (pred==2));
wanted_points_cnts= data(wanted_points,1);

detected=table(wanted_points(:),wanted_points_cnts(:));


scatter(wanted_points,wanted_points_cnts)

[Walter Roberson]

I do not see a fixed random number seed?

Reminder: kmeans() does random initialization, so you would need to set the seed before the kmeans call.

[Hamid]

@Walter Roberson Many thanks for the useful comment. I removed it here and maybe the line I wrote my randon seed was wrong.

Answer 1

[KSSV]

    
%%%%%%%%%
rng(1)
find_kmeans=kmeans(data(:,1),2);
data_2=table(data(:,1),find_kmeans(:));
data_2.Properties.VariableNames = ["counts", "detect"];

Mdl=fitcknn(data_2,'detect');
Mdl.NumNeighbors=1;

test_data=data(:,1);
pred=predict(Mdl,test_data);

wanted_points=(find (pred==2));
wanted_points_cnts= data(wanted_points,1);

detected=table(wanted_points(:),wanted_points_cnts(:));


scatter(wanted_points,wanted_points_cnts)
drawnow

[Hamid]

@KSSV Thanks a lot.