DeLorean (version 1.5.0)

estimate.hyper: Estimate hyperparameters for model using empirical Bayes.

Description

Estimate hyperparameters for model using empirical Bayes.

Usage

estimate.hyper(dl, sigma.tau = 0.5, length.scale = NULL,
  model.name = "exact", adjust.cell.sizes = TRUE)

Arguments

dl

de.lorean object

sigma.tau

Noise s.d. in temporal dimension, that is prior s.d. for tau

length.scale

Length scale for stationary GP covariance function. Defaults to the range of the observed capture times.

model.name

The model's name:

• 'exact': The model without a low rank approximation that does not estimate the cell sizes.

• 'exactsizes': The model without a low rank approximation that does estimate the cell sizes.

• 'lowrank': Low rank approximation to the 'exact' model.

• 'lowranksizes': Low rank approximation to the 'exactsizes' model.

adjust.cell.sizes

Adjust by the cell sizes for better estimates of the hyperparameters

Examples

# NOT RUN {
data(WindramDeLorean)
dl <- de.lorean(windram.expr, windram.gene.meta, windram.cell.meta)
dl <- estimate.hyper(dl)

# }