Calculate the perplexity of a celda model

Source: R/perplexity.R

Perplexity is a statistical measure of how well a probability model can predict new data. Lower perplexity indicates a better model.

perplexity(
  x,
  celdaMod,
  useAssay = "counts",
  altExpName = "featureSubset",
  newCounts = NULL
)

# S4 method for SingleCellExperiment,ANY
perplexity(
  x,
  useAssay = "counts",
  altExpName = "featureSubset",
  newCounts = NULL
)

# S4 method for ANY,celda_CG
perplexity(x, celdaMod, newCounts = NULL)

# S4 method for ANY,celda_C
perplexity(x, celdaMod, newCounts = NULL)

# S4 method for ANY,celda_G
perplexity(x, celdaMod, newCounts = NULL)

Arguments

x

Can be one of

  • A SingleCellExperiment object returned by celda_C, celda_G or celda_CG, with the matrix located in the useAssay assay slot. Rows represent features and columns represent cells.

  • Integer counts matrix. Rows represent features and columns represent cells. This matrix should be the same as the one used to generate celdaMod.

celdaMod

Celda model object. Only works if x is an integer counts matrix.

useAssay

A string specifying which assay slot to use if x is a SingleCellExperiment object. Default "counts".

altExpName

The name for the altExp slot to use. Default "featureSubset".

newCounts

A new counts matrix used to calculate perplexity. If NULL, perplexity will be calculated for the matrix in useAssay slot in x. Default NULL.

Value

Numeric. The perplexity for the provided x (and

celdaModel).

Examples

data(sceCeldaCG)
perplexity <- perplexity(sceCeldaCG)
data(celdaCGSim, celdaCGMod)
perplexity <- perplexity(celdaCGSim$counts, celdaCGMod)
data(celdaCSim, celdaCMod)
perplexity <- perplexity(celdaCSim$counts, celdaCMod)
data(celdaGSim, celdaGMod)
perplexity <- perplexity(celdaGSim$counts, celdaGMod)