0

我目前有以下功能,但是有没有更有效的方法来生成一个范围内的随机整数并在 Matlab 中排除特定整数?

function aNew = random(a)
aMin = a-100;
aMax = a+100;
aNew = a;
while aNew == a
    aNew = randi([aMin, aMax]);
end
4

4 回答 4

1

为什么不做这样的事情:

function aNew = random(a, sz)

    if nargin == 1, sz = [1 1]; end

    aMin = a-100;
    aMax = a+100;

    aNew = randi([aMin aMax-1], sz);

    aNew(aNew == a) = aMax;

end
于 2012-08-29T05:28:03.150 回答
1

您当前的方法进行拒绝抽样。你为什么不这样做:假设你要从中抽取样本的整数范围是[a, c],而你要排除的数字是b。然后[a, c-1]对每个大于或等于 的样本进行采样,并将b其递增 1。

于 2012-08-29T04:29:42.770 回答
1

您可以使用该函数randsample从指定分布中抽取样本。

像这样的东西应该可以正常工作

sampleRange = [1, 100];  % sample from 1 to 100
noSample = 50; % lets exclude 50
pop2Sample = [range(1):noSample-1, noSample+1:range(2)]; %create the population

sample = randsample(pop2Sample,1); %draw a single sample

更新

如果您想从总体中排除多个值,您可以使用该setdiff函数。

pop2Sample = 1:100;  % sample from 1 to 100
noSample = 0:10:100 % lets exclude any all multiples of ten
pop2Sample = setdiff(pop2Sample, noSample);

sample = randsample(pop2Sample,1); %draw a single sample
于 2012-08-29T02:20:47.217 回答
0

按照@slayton 的解决方案,为了对样本进行统一抽样N,有放回,您可以在值总体内随机索引,有放回:

N = 10; % samples
aMin = -100; % lower
aMax = +100; % upper
a = 0; % excluded int 

valPop = [aMin:a-1, a+1:aMax]; % value population
samples = valPop(randi([1, N-1], 1, N)); % with resampling
于 2012-08-29T21:24:13.340 回答