Is there a way to holdout specific data?

Question

Mark 2024-9-23，10:48

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2154810-is-there-a-way-to-holdout-specific-data

评论： Udit06 2024-9-23，15:07

I'm producing decision trees (both classification and regression) of my dataset and I wish to use a specific set of data as the training and a specific set for the testing. Is there a way to do this?

For example, say my dataset is consists of 100 rows, is there a way to tell the software to compute rows 1-75 as the training set and rows 76-100 as the test set?

Thanks in advance

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

Udit06 2024-9-23，10:57

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2154810-is-there-a-way-to-holdout-specific-data#answer_1521205

在 MATLAB Online 中打开

Hi Mark,

You can use the array indexing to specify the training and testing sets using indexes. Please find below the code snippet to achieve the same:

% Define the training and testing indices
trainIndices = 1:75;
testIndices = 76:100;
% Split the data
trainData = data(trainIndices, :);
testData = data(testIndices, :);

I hope this helps.

2 个评论
显示无隐藏无

Mark 2024-9-23，13:06

Thanks for you help. Apologies if this next questions is a bit dumb, I'm still getting to grips with Matlab, but how do I then add that in to the code for computing the tree? For example, I use the below code:

MdlA = fitctree(input,output); % for unvalidated trees

MdlB = fitctree(input,output,'CrossVal','on') % for validated trees.

How in the case of MdlB, do I basically tell it to use rows 1-75 for the initial output, but 76-100 for the cross validation step?

Udit06 2024-9-23，15:07

在 MATLAB Online 中打开

Hi Mark,

When you use the fitctree function with the 'CrossVal','on' option, MATLAB automatically performs cross-validation by splitting the data into multiple folds. You can find the same on the following MathWorks documentation:

https://www.mathworks.com/help/stats/fitctree.html#:~:text=Train%20a%20classification%20tree%20using%20the%20default%20values%20for%20tree%20depth%20control.%20Cross%2Dvalidate%20the%20model%20by%20using%2010%2Dfold%20cross%2Dvalidation.

However, if you want to manually specify the training and test sets, you should handle the splitting yourself rather than relying on the built-in cross-validation. You can find the code snippet on how to train the model using manually specifying train and test sets:

% Clear workspace

clear;

% Load the ionosphere dataset

load ionosphere;

% Define training data ratio and calculate number of training samples

trainRatio = 0.7;

numTrainSamples = round(trainRatio * size(X, 1));

% Split data into training and test sets

X_train = X(1:numTrainSamples, :);

Y_train = Y(1:numTrainSamples);

X_test = X(numTrainSamples+1:end, :);

Y_test = Y(numTrainSamples+1:end, :);

% Train a decision tree classifier

MdlA = fitctree(X_train, Y_train);

% Visualize the decision tree

view(MdlA, 'Mode', 'graph');

% Predict labels for the test set

Y_pred = predict(MdlA, X_test);

% Convert cell arrays to matrices for comparison

Y_test = cell2mat(Y_test);

Y_pred = cell2mat(Y_pred);

% Calculate and display accuracy

accuracy = sum(Y_test == Y_pred) / length(Y_test);

fprintf('Test Set Accuracy: %.2f%%\n', accuracy * 100);

Test Set Accuracy: 89.52%

I hope this helps.

请先登录，再进行评论。

Is there a way to holdout specific data?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

回答（1 个）

2 个评论
显示无隐藏无

另请参阅

类别

标签

产品

版本

Community Treasure Hunt

Is there a way to holdout specific data?

0 个评论 显示 -2更早的评论隐藏 -2更早的评论

回答（1 个）

2 个评论 显示 无隐藏 无

另请参阅

类别

标签

产品

版本

Community Treasure Hunt

WeChat

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

2 个评论
显示无隐藏无