如何产生与几率相关矩阵

lpp8515

matlab里有没这样的函数？或者可以编程实现以下功能，比如给出一个一维矩阵A，其值介于0－－100之间，把这个矩阵二值化为只有0和1两个值的矩阵B，方法为：假设A=[97    79    50    21     3     3    21    50    79    97];我想让生成的矩阵B中的元素个数与矩阵A中的元素个数同（size(A)=size(B)），当A(1)=97时，B(1)有97%的几率等于1，当A(2)=79时，B(2)有79%的几率等于1(即有21%的几率等于0)，以此类推。最后生成的B为1*10的一维矩阵，且只包含0和1两个数值。这个怎么实现？还望高手指点。

[ 本帖最后由 ChaChing 于 2010-4-18 11:32 编辑 ]

yufeng

A=[97    79    50    21     3     3    21    50    79    97];
c=rand(1,10)*100;B=A>c

ChaChing

请扫盲下
rand出来的数字, 代表几率吗?

xiezhh

试试这个：

1. 
   
2. >> A = [97 79  50  21  3  3  21  50  79  97];
   
3. >> R = mnrnd(1,[A'/100 1-A'/100]);
   
4. >> B = R(:,1)'
   
5. B =
   
6.      1     1     1     0     0     0     0     1     0     1
   

复制代码

ChaChing

现在的版本较旧(无mnrnd函数), 尚无法验证学习!
但谢老师的, 个人有信心, 先加威望了

xiezhh

谢谢ChaChing大哥的信任和鼓励！

maigicku

你好，谢老师，如果我想生成一个100X100的0 1矩阵，其中只有100个1，就是说1的概率是0.1，怎么生成呢？

dreamstone

研究下这个就该差不多了吧
mnrnd

Multinomial random numbers
Syntax

r = mnrnd(n,p)
R = mnrnd(n,p,m)
R = mnrnd(N,P)
Description

r = mnrnd(n,p) returns random values r from the multinomial distribution with parameters n and p. n is a positive integer specifying the number of trials (sample size) for each multinomial outcome. p is a 1-by-k vector of multinomial probabilities, where k is the number of multinomial bins or categories. p must sum to one. (If p does not sum to one, r consists entirely of NaN values.) r is a 1-by-k vector, containing counts for each of the k multinomial bins.

R = mnrnd(n,p,m) returns m random vectors from the multinomial distribution with parameters n and p. R is a m-by-k matrix, where k is the number of multinomial bins or categories. Each row of R corresponds to one multinomial outcome.

R = mnrnd(N,P) generates outcomes from different multinomial distributions. P is a m-by-k matrix, where k is the number of multinomial bins or categories and each of the m rows contains a different set of multinomial probabilities. Each row of P must sum to one. (If any row of P does not sum to one, the corresponding row of R consists entirely of NaN values.) N is a m-by-1 vector of positive integers or a single positive integer (replicated by mnrnd to a m-by-1 vector). R is a m-by-k matrix. Each row of R is generated using the corresponding rows of N and P.
Example

Generate 2 random vectors with the same probabilities:

n = 1e3;
p = [0.2,0.3,0.5];
R = mnrnd(n,p,2)
R =
   215   282   503
   194   303   503

Generate 2 random vectors with different probabilities:

n = 1e3;
P = [0.2, 0.3, 0.5; ...
     0.3, 0.4, 0.3;];
R = mnrnd(n,P)
R =
   186   290   524
   290   389   321

ChaChing

ls给的就是matlab的help嘛!?

maigicku

已经解决了，用randperm