i have
dataset=[6 7; 5 4; 9 8; 1 2; 9 8; 4 5; 1 2; 3 4; 8 7; 6 2] can random select 90% of data training , remaining (10%) test set repeat split 10 times.
i.e
training = [6 7; 5 4; 9 8; 1 2; 9 8; 1 2; 3 4; 8 7; 6 2] test= [4 5] i wrote code
num_points = size(x,2); split_point = round(num_points*0.7); to split data can't obtain result
dataset=[6 7; 5 4; 9 8; 1 2; 9 8; 4 5; 1 2; 3 4; 8 7; 6 2] randomly re-order dataset using randperm:
n = size(dataset,1); data_rand = dataset(randperm(n),:) then pull out different 10% each time:
m = ceil(n/10); group = 1; k = 1:m:n-m test{group} = data_rand(k:k+m-1,:) train{group} = [data_rand(1:k-1,:); data_rand(k+m:end,:)]; group = group + 1; end but suggest read cross validation in matlab has lot of built-in functionality this.
Comments
Post a Comment