How to Split fisher iris data into 60% training and 40% Testing

Question

hammad younas 2021-12-6

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1603910-how-to-split-fisher-iris-data-into-60-training-and-40-testing

回答： yanqi liu 2021-12-7

在 MATLAB Online 中打开

Hello I hope you are doing well.

I want to split the fisher iris dataset betwee 60% training and 40% testing Dataset How can i divide that?

i am using this Example

https://www.mathworks.com/help/stats/classification-example.html

It used all training examples not test example i want to divide it betwee train and test

load fisheriris
f = figure;
gscatter(meas(:,1), meas(:,2), species,'rgb','osd');
xlabel('Sepal length');
ylabel('Sepal width');
N = size(meas,1);
lda = fitcdiscr(meas(:,1:2),species);
ldaClass = resubPredict(lda);

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Chunru 2021-12-6

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1603910-how-to-split-fisher-iris-data-into-60-training-and-40-testing#answer_848420

编辑：Chunru 2021-12-6

在 MATLAB Online 中打开

load fisheriris

n = size(meas, 1);

%hpartition = cvpartition(n, 'holdout', 0.4); % 40% for test

hpartition = cvpartition(species, 'holdout', 0.4); % 40% for test

idxTrain = training(hpartition);

idxTest = test(hpartition);

pie(categorical(species(idxTrain))); % distribution of training samples

XTrain = meas(idxTrain, :);

TTrain = species(idxTrain);

XTest = meas(idxTest, :);

TTest = species(idxTest);

% Training

lda = fitcdiscr(XTrain(:,1:2), TTrain);

% Prediction

testClass = predict(lda, XTest(:, 1:2));

4 个评论
显示 2更早的评论隐藏 2更早的评论

hammad younas 2021-12-6

@Chunru it does not divide the dataset equaly like 20 for each class

Chunru 2021-12-6

在 MATLAB Online 中打开

For approximately equal partition:

hpartition = cvpartition(species, 'holdout', 0.4);

请先登录，再进行评论。

Answer 2

yanqi liu 2021-12-7

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1603910-how-to-split-fisher-iris-data-into-60-training-and-40-testing#answer_849170

在 MATLAB Online 中打开

yes，sir，may be use the follow split method ，such as

close all;
clear all;
clc;
load fisheriris
cs = categorical(species);
ds = categories(cs);
training_x = [];training_y = [];
testing_x = [];testing_y = [];
for i = 1 : length(ds)
    ind = find(cs == ds{i});
    % rand suffer
    ind = ind(randperm(length(ind)));
    %  60% training and 40% testing
    training_x = [training_x; meas(ind(1:round(length(ind)*0.6)),:)];
    training_y = [training_y; cs(ind(1:round(length(ind)*0.6)),:)];
    testing_x = [testing_x; meas(ind(1+round(length(ind)*0.6):end),:)];
    testing_y = [testing_y; cs(ind(1+round(length(ind)*0.6):end),:)];
end