NormMixParam {coseq} | R Documentation |
Calculates the mean and covariance parameters for a normal mixture model of the form pK_Lk_Ck
NormMixParam( coseqResults, y_profiles = NULL, K = NULL, digits = 3, plot = FALSE, ... )
coseqResults |
Object of class |
y_profiles |
y (n x q) matrix of observed profiles for n
observations and q variables, required for |
K |
The model used for parameter estimation for objects |
digits |
Integer indicating the number of decimal places to be used for output |
plot |
If |
... |
Additional optional parameters to pass to |
pi |
Vector of dimension K with the estimated cluster proportions from the Gaussian mixture model, where K is the number of clusters |
mu |
Matrix of dimension K x d containing the estimated mean
vector from the Gaussian mixture model, where d is the
number of samples in the data |
Sigma |
Array of dimension d x d x K containing the
estimated covariance matrices from the Gaussian mixture model, where d is the
number of samples in the data |
rho |
Array of dimension d x d x K containing the
estimated correlation matrices from the Gaussian mixture model, where d is the
number of samples in the data |
Andrea Rau, Cathy Maugis-Rabusseau
## Simulate toy data, n = 300 observations set.seed(12345) countmat <- matrix(runif(300*4, min=0, max=500), nrow=300, ncol=4) countmat <- countmat[which(rowSums(countmat) > 0),] profiles <- transformRNAseq(countmat, norm="none", transformation="arcsin")$tcounts conds <- rep(c("A","B","C","D"), each=2) ## Run the Normal mixture model for K = 2,3 ## Object of class coseqResults run <- NormMixClus(y=profiles, K=2:3, iter=5) run ## Run the Normal mixture model for K=2 ## Object of class SummarizedExperiment0 run2 <- NormMixClusK(y=profiles, K=2, iter=5) ## Summary of results summary(run) ## Re-estimate mixture parameters for the model with K=2 clusters param <- NormMixParam(run, y_profiles=profiles)