Split array into training and testing

Hi,

I have a set of data (DataA has 106x14). I want to split the rows into 2 section, one for training and one for testing. Here is my code:

[trainA,testA] = divideblock(DataA.', .7, .3); % 70% for training 30% for testing.

trainData = trainA.'; 
testData = testA.';

Result: but the total data I have after executing the code is only 93 (66x14 for traindata, 27x14 for testdata) I don't want to use valInd since I don't need it.

PLEASE correct me.

[Cyrus]

I would manually find the 70 and 30 % and use "for loop" to split the data

Answer 1

[Jos (10584)]

dataA = cumsum(ones(20,3))  % some test data
p = .7      % proportion of rows to select for training

N = size(dataA,1)  % total number of rows 
tf = false(N,1)    % create logical index vector
tf(1:round(p*N)) = true     
tf = tf(randperm(N))   % randomise order
dataTraining = dataA(tf,:) 
dataTesting = dataA(~tf,:)

[Ihsan Yassin]

Thanks, it works exactly like I want. but may I ask another question,

tf = false(N,1)    % create logical index vector
tf(1:round(p*N)) = true     
tf = tf(randperm(N))   % randomise order

what are these for, I don't quite understand. Thanks again for your help.

Answer 2

[Greg Heath]

Your answer should be simply obtained from the divideblock documentation (help and/or doc). From the help documentation example

>> clear all, help divideblock

    [trainInd,valInd,testInd] = divideblock(250,0.7,0.15,0.15);
    whos

    Name          Size     Bytes  Class

   testInd       1x37       296  double               
   trainInd      1x176     1408  double              
   valInd        1x37       296  double

>> [ 176 37 37 ]/250

ans =

    0.7040    0.1480    0.1480

However, DIVIDEBLOCK (MATLAB 2016A) HAS A BUG

>> clear all, clc

 [trainInd,valInd,testInd] = divideblock(250,0.7,0.0,0.3);  
 whos

 Subscript indices must either be real positive integers or logicals

 Error in divideblock>divide_indices (line 108)
  testInd = (1:numTest)+valInd(end);

 Error in divideblock (line 65)
    [out1,out2,out3] = divide_indices(in1,params);

Hope this helps.

Thank you for formally accepting my answer

Greg

P.S. What version of MATLAB do you have?

[Ihsan Yassin]

I have 2015 ver..yes I get those error when i wrote that, as I don't want to use that valInd

Answer 3

[Jaeseok Kim]

dividerand is what you want...

Answer 4

[Satyam Agarwal]

[Trainset,Testset]= splitEachLabel(datastore,p)

p is ratio 0<p<1

Answer 5

[MUHAMMAD SAJAD]

% Split 60% of the files from each label into ds60 and the rest into dsRest [ds60,dsRest] = splitEachLabel(imds,0.6) ds60 is a trainingset while dsRest is testset. we can also divide it for validset. like this [TrianSet,ValidSet,TestSet]=splitEachLabel(DataStore,0.7,0.2). In this case 70% of files split for TrainingSet,20% for ValidSet and the remaining for TestSet.

Answer 6

[indhumathi karuppaiya]

hi my name indhu i try to do project for my studies .i have choosed parkinson diease speech recognition in matlab coding how to split the data to train data and test data please let me know just i want use only 1to 60 patiend data onlu use thank u