Randomly split data set and make sure that a sample of each class is included

4 次查看（过去 30 天）
CHRISTOPHER MILLAR
评论： CHRISTOPHER MILLAR 2020-1-8
Hello,
I am randomly splitting my dataset into train,validation and test sets as shown below:
% Random split Complete Data Set
Q = size(completeRDS,1);
Q1 = floor(Q*0.75);
Q2 = Q-Q1;
Q2 = round(Q2/2);
Q3 = Q2;
ind = randperm(Q);
ind1 = ind(1:Q1);
ind2 = ind(Q1+(1:Q2));
ind3 = ind(Q2+(1:Q3));
% Complete Data Set
trainData = completeRDS(ind1,1);
trainSeqLabels = seqLabels(ind1,1);
valData = completeRDS(ind3,1);
valLabels = seqLabels(ind3,1);
testData = completeRDS(ind2,1);
testLabels = seqLabels(ind2,1);
I have just realised that my test set has not included a sample from each class that I am trying to classify. It has not included 2 of the classes
How can i split my data set and make sure that the test set includes at least 1 sample of each class that i am trying to classify?
Thanks in advance

请先登录，再回答此问题。

回答（1 个）

MaryD
MaryD 2020-1-8
If those are images I think best way is to load your data with class labels and then use splitEachLabel( __,'randomize') function.

类别

在 Help Center 和 File Exchange 中查找有关 Image Data Workflows 的更多信息

标签

产品


版本

R2019b

Community Treasure Hunt

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

Start Hunting!

Translated by