PSVDOS: Poisson SVD with Offset estimation

In seonjoo/PSVDOS: Conduct poisson singular value decomposition with offset estimation

Description

Poisson SVD with Offset estimation

Usage

PSVDOS(Y = NULL, K = NULL, B = NULL, F = NULL, E = NULL,
  niter = 100, err = 1e-04, ln = "log", verbose = 0, const = 1,
  zerocount = exp(-3), rowoffset = NULL, colcenter = 0, ...)

Arguments

Y:	data matrix, n-by-m matrix.
B:	the first K columns of the U from SVD(log(Y/T)), n
F:	the first K columns of the SV from SVD(log(Y/T))
K:	# of SVD components
niter:	# of maximum iterations for poissonSVD
err:	error bound on claiming convergence
ln:	link functions, "log", "sqrt", "identity"
verbose:	Log information

Value

Returns PSVD results

B: Estimated

References

[1] Lee, S., Chugh, P. E., Shen, H., Eberle, R., & Dittmer, D. P. (2013) Poisson factor models with applications to non-normalized microRNA profiling. Bioinformatics, 29(9), 1105-1111

Examples

data(demo)
heatmap(Y)
psvdfit = PSVDOS(Y,K=4,verbose=1,err = 0.0001,niter = 300)
names(psvdfit)

seonjoo/PSVDOS documentation built on May 29, 2019, 6:54 p.m.