run_em: Run EM to estimate parameters of mixture model
In marcalva/diem: Debris-Containing Droplet Identification using EM

run_em

R Documentation

Run EM to estimate parameters of mixture model

Description

Estimate the parameters of the mixture model, and estimate the posterior probability a droplet belongs to each of the clusters.

Usage

run_em(x, A = NULL, P = NULL, Z = NULL, model = "mltn",
  alpha_prior = 0, pi_prior = 0, max_iter = 1000, eps = 1e-04,
  psc = 1e-10, threads = 1, verbose = TRUE)

Arguments

`x`	An SCE object.
`A`	Initial alpha parameter.
`P`	Initial pi parameter.
`Z`	Initial latent variable Z values.
`model`	The mixture model to assume. Can be either "DM" for a Dirichlet-multinomial or "mltn" for a multinomial.
`alpha_prior`	Add a non-informative prior by adding a count of `alpha_prior` to all genes in each cluster. Only valid for the multinomial model.
`pi_prior`	Add a non-informative prior by adding a count of `pi_prior` to the each cluster's membership.
`max_iter`	Maximum number of iterations for the EM estimation of the mixture model.
`eps`	The delta threshold for when to call convergence for the EM estimation of the mixture model. The EM stops when delta falls below this value. We define delta as the average change in posterior probability. By default this is set to 1e4, so that the EM converges when less than 1 in 10,000 labels change on average.
`psc`	Pseudocount to add to avoid collapsing likelihood to 0.
`threads`	Number of threads for parallel execution. Default is 1.
`verbose`	Verbosity.