general.fi.samplesize: Sample size calculator, taking into account power and...
In brb225/FragilityTools: Toolbox For Patient Count Fragility Measures

Description Usage Arguments Value Examples

This function takes in a function to compute a p value and a function to simulate data. Using these, it finds the smallest sample size which produces a power and fragility index larger than the input thresholds.

general.fi.samplesize(
  min.fi = 10,
  min.power = 0.8,
  sample_size_init_power = 100L,
  sample_size_init_fi = NULL,
  get.p.val,
  get.replacements,
  get.sample,
  gamma = 0.6,
  niters = 50,
  cl = NULL,
  verbose = FALSE,
  alpha = 0.05,
  tau = 1/2,
  nsim = 30,
  eps = 0.1,
  algorithm = "walsh"
)

`min.fi`	the smallest acceptable QUANTILE fragility index. When NULL, the FI calculation is skipped and sample_size_init_fi is taken to produce the desired FI.
`min.power`	the smallest acceptable power. When NULL, the power calculation is skipped and sample_size_init_power is taken to produce the desired power.
`sample_size_init_power`	a sample size to initialize the algorithm (not necessary, defaults to 10) to find the sample size for power
`sample_size_init_fi`	a sample size to initialize the algorithm (not necessary, defaults to the sample size for power) to find the sample size for fi
`get.p.val`	a fucntion that inputs a matrix and returns a p value, otherwise a function that inputs X and Y and returns a p value when alg='greedy'.
`get.replacements`	a function which outputs a data frame containing all possible row replacements for Y which are to be considered. The functions inputs the row of Y under consideration, the row of X under consideration, the row name, and the full original data frames X and Y. This gets used in the greedy algorithm.
`get.sample`	a function which inputs a sample size and outputs a sample, with the covariates in a dataframe X and the response in a dataframe Y
`gamma`	the power of n^-1 in the gradient descent in the Polyak-Ruppert averaging. The default is 0.60.
`niters`	The number of times to repeat the sample size calculations. The median of the repitions is then reported.
`cl`	a cluster from the `parallel` package, used for the loop over 1:niters
`verbose`	boolean to indicate whether to print out algorithmic information for the sample size calculation
`alpha`	a number for the size of test
`tau`	the quantile of FI to bound, default 1/2
`nsim`	the number of simulated draws to consider when estimating the power and fragility index quantile
`eps`	a parameter to control the error. The smaller is it, the more precise the output but the longer the function will take to run.
`algorithm`	A string specifying the algorithm to use to calculate fragility indices. The default is "walsh". Alternatives include "greedy" for using greedy.fi

the length two numeric vector of the calculated sample sizes for the desired power and fragility index

ss <- general.fi.samplesize(min.fi = 10, min.power = .8, alpha = .05, tau=.5,
get.p.val = function(m)fisher.test(m)$p.value, get.sample = function(ss) draw.binom(ss, matrix=TRUE),
nsim = 10, niters = 2, verbose = TRUE)

ss <- general.fi.samplesize(min.fi = 10, min.power = .8, alpha = .05, tau=.5,
get.p.val = function(X,Y) fisher.test(table(X[,1], Y[,1]))$p.value,
get.sample = function(ss) draw.binom(ss, matrix=FALSE),
alg='greedy', niters=1, verbose = TRUE,
get.replacements=function(y, x, rn, Y, X) data.frame(setdiff(c("event", "nonevent"), y)))