how to partition a dataset into 60% v 40%

Question

sam 2014-11-9

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/161970-how-to-partition-a-dataset-into-60-v-40

评论： the cyclist 2020-2-13

I have a diabaetes.mat file downloaded. I want to partition the data set into two groups 60% training set and 40% test set. I then want to rank the features.

I figure to rank the features i will use the function corrcoef but i have no idea how to partian the data set into 60% vs 40%.

cheers sam

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

the cyclist 2014-11-9

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/161970-how-to-partition-a-dataset-into-60-v-40#answer_158279

在 MATLAB Online 中打开

Depending on what toolboxes you have installed, there are a number of options:

cvpartition
randsample
randperm

The first two require the Statistics Toolbox, but that last one is in core MATLAB.

2 个评论
显示无隐藏无

Taiwo Kupoluyi 2020-2-13

编辑：Taiwo Kupoluyi 2020-2-13

I guess the question to ask(for smeone new to Matlab) is what Toolbox do i need to have in order to be able to partition a dataset into Training and Holdout data.

Thank you in anticipation of your response.

the cyclist 2020-2-13

You don't have to have any Toolbox to partition a dataset. You can partitiion a dataset into training and holdout using the randperm function (in base MATLAB) to randomly order the data, and then pick the first 80% (for example) for training.

But the cvpartition and randsample functions might make the job a little easier. Also, the Statistics and Machine Learning Toolbox is likely to have many other functions you might want to use for modeling.

This is kind of a general rule for toolboxes. You could write everything from scratch if you want to. Getting a toolbox is paying for the convenience (and rigor) of having MathWorks do it.

请先登录，再进行评论。