fit.mixture.model: Fit a Gaussian mixture deconvolution model

In qingyuanzhao/mr.raps: Two Sample Mendelian Randomization using Robust Adjusted Profile Score

Fit a Gaussian mixture deconvolution model

Description

Fit a Gaussian mixture deconvolution model

Usage

fit.mixture.model(
  z,
  n = 2,
  ntry = 20,
  force.mu.zero = TRUE,
  diagnostics = FALSE
)

Arguments

z	A vector of z-scores.
n	Number of mixture components.
ntry	Number of random initializations.
force.mu.zero	Should the means be forced to zero?
diagnostics	Logical indicator for showing diagnostic plots.

Details

This function assumes that z is distributed as N(γ, 1) and γ follows a Gaussian mixture model. It fits this deconvolution model by maximum likelihood and outputs the estimated mixture distribution.

Value

A list of p (mixture proportion), mu (mean), sigma (standard deviation).

Examples

z <- c(sqrt(2) * rnorm(900), sqrt(17) * rnorm(100))
## So the correct sigma = (1, 4) and p = (0.9, 0.1)
fit.mixture.model(z)

qingyuanzhao/mr.raps documentation built on June 4, 2022, 3:04 a.m.