pkgdown/in_header.html

Skip to contents

MultinomialCGF

Source: R/C-MultinomialCGF.R
MultinomialCGF.Rd

A ready-to-use CGF object for handling multiple i.i.d. multinomial random vectors. The number of multinomial random vectors is determined by the length of tvec and the length of the multinomial probability vector. For user-specified parameter mappings, refer to MultinomialModelCGF.

Usage

MultinomialCGF

Format

An object of class CGF (an R6 class), with standard methods K, K1, K2, K3operator, K4operator, etc.

Details

Parameter Vector Interpretation: The parameter_vector used when calling methods like K1(tvec, parameter_vector) should be of the form \((n, x_1, \dots, x_d)\), where:

  • \(n\) is the total count (a positive integer),

  • \(x = (x_1, \dots, x_d)\) is a vector of non-negative entries, not all zero.

The vector \(x\) can be interpreted in two ways:

  • Odds: \(x_i\) are odds values. Probabilities are derived as \(p_i = x_i / \sum_{j=1}^{d} x_j\).

  • Probabilities: If \(x\) sums to 1, it is treated directly as the probability vector \(p\).

IID Assumption: All multinomial random vectors processed by MultinomialCGF are assumed to be i.i.d.

Examples

tvec <- rep(0,3)
parameter_vector <- c(10, 2, 3, 5)
# total count=10, odds=(2,3,5) or probabilities=c(0.2,0.3,0.5)
MultinomialCGF$K1(tvec, parameter_vector)  # MultinomialCGF$K1(tvec, c(0.2,0.3,0.5))
#> [1] 2 3 5