# rmulti: Random draws from a multinomial distribution In poLCA: Polytomous variable Latent Class Analysis

## Description

One random draw from a multinomial distribution or list of multinomial distributions.

## Usage

 `1` ```rmulti(p) ```

## Arguments

 `p` matrix of dimension `n` by `r` containing probabilities, for each row, of drawing each of `r` outcomes. `p` may also be entered as a vector, in which case `rmulti` treats it as a matrix of dimension `n=1` by `r`.

## Value

Returns a vector of length `n`. Each item represents one draw from the multinomial distribution parameterized by the outcome probabilities in each row of `p`.

## Note

Each row of matrix `p` must sum to 1 or `rmulti` will not work properly.

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28``` ```## ## One draw from a three-category multinomial distribution. ## p1 <- c(0.7,0.2,0.1) rmulti(p1) ## ## 10,000 draws from a three-category multinomial distribution. ## n <- 10000 p2 <- matrix(p1,nrow=n,ncol=length(p1),byrow=TRUE) rmdraws <- rmulti(p2) table(rmdraws)/n # should be approximately 0.7, 0.2, 0.1 ## ## 10,000 draws from a mixture of three groups of a ## four-category multinomial distribution. ## group.p <- matrix(c(0.5,0.3,0.2),nrow=n,ncol=3,byrow=TRUE) group <- rmulti(group.p) p3 <- t(matrix(NA,nrow=n,ncol=4)) p3[,group==1] <- c(0.7,0.1,0.1,0.1) p3[,group==2] <- c(0.1,0.7,0.1,0.1) p3[,group==3] <- c(0.1,0.1,0.1,0.7) p3 <- t(p3) rmdraws3 <- rmulti(p3) table(group,rmdraws3) table(group,rmdraws3)/rowSums(table(group,rmdraws3)) ```

### Example output

```Loading required package: scatterplot3d
[1] 3
rmdraws
1      2      3
0.6992 0.1989 0.1019
rmdraws3
group    1    2    3    4
1 3469  481  506  525
2  297 2106  297  316
3  211  177  189 1426
rmdraws3
group          1          2          3          4
1 0.69644650 0.09656695 0.10158603 0.10540052
2 0.09847480 0.69827586 0.09847480 0.10477454
3 0.10534199 0.08836745 0.09435846 0.71193210
```

poLCA documentation built on May 29, 2017, 5:59 p.m.