Simulate multinomial data under the null (strong or weak) or alternative
simulate_data_mult.RdBy default, \(p = 2\) with a single Uniform(0,1) continuous covariate.
Usage
simulate_data_mult(
nn,
null = TRUE,
strong = FALSE,
alt_magnitude = 1,
jj = 5,
ms = 10000,
jj_null = NULL,
Beta = NULL,
sd_beta0s = NULL,
sd_beta1s = NULL,
overdispersion = 0,
covariate = NULL
)Arguments
- nn
Number of observations
- null
If TRUE will simulate under the null, if FALSE will simulate under the alternative
- strong
If null is TRUE, simulate under the strong null? Defaults to FALSE (simulate under the weak null)
- alt_magnitude
The mean of each parameter in the beta1 vector if null = FALSE. Defaults to \(1\).
- jj
Number of taxa
- ms
Number of counts per sample
- jj_null
For the weak null, which taxon should be null
- Beta
User-specified value of the true beta (if you wish to draw from a fixed beta rather than generate beta0's and beta1's).
- sd_beta0s
The beta0's are drawn from a normal distribution with mean zero. This is the standard deviation of that distribution.
- sd_beta1s
The beta1's are drawn from a normal distribution with mean zero under the null or non-zero under the alternative. This is the standard deviation of that distribution.
- overdispersion
An additional normal random variable can be added to the link function to add dispersion above a multinomial distribution. This is the standard aviation for that normal variable. Useful for confirming error rate control under model misspecification.
- covariate
An optional covariate vector, if not provided the covariate will be a sequence of increasing values from 0 to 1.