Get_topics: Obtaining the result by performing topic model.

In BinDuan/MASCOT: Model-based Analysis of Single cell CRISPR knockOuT screening

Description

Obtaining the result by performing topic model with selection of the optimal topic number automatically.

Usage

Get_topics(expression_profile_var_gene, sample_info_gene, topic_number_min = 3, topic_number_max = 8, alpha = 0.5, seed_num = 2017, burnin = 1000, thin = 100, iter = 1000, plot = FALSE, plot_path = "~/select_topic_number.pdf")

Arguments

expression_profile_var_gene	A dataframe showing the expression profile only for the selected highly dispersion differentially expressed genes.
sample_info_gene	A character vector showing the knockout gene of each sample after all the filterings.
topic_number_min	The minimum candidate topic number.
topic_number_max	The maximum candidate topic number.
alpha	alpha is the weight of specificity score. Its value is between 0 and 1. The default is 0.5.
plot	FALSE by default. If TRUE, plot the graph.
plot_path	The path of the graph you plot. It works only when the parameter "plot" is TRUE.
seed_num	Object of class "integer"; used to set the seed for Gibbs sampling. Default 2017.
burnin	Object of class "integer"; number of omitted Gibbs iterations at beginning, by default 1000.
thin	Object of class "integer"; number of omitted in-between Gibbs iterations, by default equals iter.
iter	Object of class "integer"; number of Gibbs iterations, by default equals 1000.

Value

Object of class "LDA" with the optimal topic number.

Author(s)

Bin Duan

References

Blei D.M., Ng A.Y., Jordan M.I. (2003). Latent Dirichlet Allocation. Journal of Machine Learning Research, 3, 993<e2><80><93>1022.

BinDuan/MASCOT documentation built on May 23, 2019, 2:42 p.m.