plda_choose_lambda: pLDA Choose Shrinkage Parameter

In wuxiaotiankevin/pLDA: The pLDA Package

Description

Run pLDA under different lambda values. Plot and return output. Interesting genes are defined as those with variance in beta matrix greater than average variance of housekeeping genes.

Usage

plda_choose_lambda(dat, k, lam_values, idx_hk, njobs = 1, n_rep = 1,
  fdir = NA, seed = 2019)

Arguments

dat	Input matrix. A matrix of positive integers where rows represent genes (words) and columns represent cells (documents).
k	Number of topics.
lam_values	A vector of shrinkage parameter values.
idx_hk	Index (column number) of housekeeping gene.
njobs	Number of cores used for the calculation. Default 1.
n_rep	Number of repetition for each lam_values. Default 1.
fdir	Output file directory. Default ../output/
seed	LDA seed. Default 2019. If a vector, use the best seed in terms of likelihood.

Value

Plot number of interesting genes for different lambda values. Return table of lam_values and corresponding number of interesting genes.

Examples

## Not run: 
plda_choose_lambda(dat=cell_by_gene_expr_matrix, k=10, lam_values=c(10, 10^2, 10^3),  idx_hk = 1:10)

## End(Not run)

wuxiaotiankevin/pLDA documentation built on Nov. 11, 2019, 11:01 p.m.