randomly sampling with replacement from a sample.

Question

Paraic Hickey 2014-4-3

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/124411-randomly-sampling-with-replacement-from-a-sample

回答： Jos (10584) 2014-4-4

采纳的回答： dpb

Hi,

I am simulating traffic data on a highway bridge.

I want to generate 1000 trucks a day for a total of 2500 days and find the max for each day.

This means sampling with replacement 1000 trucks from the 113,000 in the data 2500 times and finding the max each time.

As far as I am aware I cannot change the size of a bootstrap sample so I have being trying to use datasample.

The code I have tried is:

myStatistic=@max

nreps=2500

y=datasample(data,1000);

for i=1:nreps; populationStat(i)=myStatistic(data(;,i));

This does not work. What code or method should I be using.

Thanks in advance

2 个评论
显示无隐藏无

Jan 2014-4-4

"Does not work" is not useful to explain a problem. What happens? Do you get an error message or do the results differ from your expectations?

Paraic Hickey 2014-4-4

I got an error message : index exceeds the matrix size. I think I have it working now though. Thanks for your help

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

dpb 2014-4-3

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/124411-randomly-sampling-with-replacement-from-a-sample#answer_132018

...sampling with replacement 1000 trucks from the 113,000 in the data...

y=truckdata(randi(113000,1000,1));

2 个评论
显示无隐藏无

Paraic Hickey 2014-4-3

Thanks

how do I loop this to create 2500 samples of 1000 and find the max of each sample set?

Joseph Cheng 2014-4-3

probably reading up on randi. reading the documentation on randi will give you an idea that you can go randi(maxpossiblenumber,2500,1000). then use max(answer,[],2) to get the max of each row.

请先登录，再进行评论。

Answer 2

Jos (10584) 2014-4-4

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/124411-randomly-sampling-with-replacement-from-a-sample#answer_132153

在 MATLAB Online 中打开

You can use BOOTSTRP. An example:

% data
  Data = 1:20 ;  % 113,000 values in your data
  Nreps = 5      % 2500 in your case
  Nselected = 3  % 1000 in your case
  FH = @(X) max(X(1:Nselected)) ; % statistics function, max of first values
% engine
  popStat = bootstrp(Nreps, FH, Data)