Calculate standard errors of clustering parameters.

Calculate SE of parameters fitted using clustord.

Usage

calc_SE_rowcluster(long_df, clust_out, control_optim = default_control_optim())

calc_SE_bicluster(long_df, clust_out, control_optim = default_control_optim())

Arguments

long_df

The data frame, in long format, as passed to clustord.

clust_out

A clustord object.

control_optim

control list for the optim call within the M step of the EM algorithm. See the control list Details in the optim manual for more info.

Value

The standard errors corresponding to the elements of clust_out$out_parvec.

Details

Use calc_SE_rowcluster to calculate SE for row clustering and column clustering, or calc_SE_bicluster to calculate SE for biclustering.

Calculates SE by running optimHess (see optim) on the incomplete-data log-likelihood to find the hessian at the fitted parameter values from clustord. Then the square roots of the diagonal elements of the negative inverse of the hessian are the standard errors of the parameters i.e. SE <- sqrt(diag(solve(-optim_hess)).

Note that SE values are only calculated for the independent parameters. For example, if the constraint on the row clustering parameters is set to constraint_sum_zero = TRUE, where the last row clustering parameter is the negative sum of the other parameters, SE values will only be calculated for the first RG-1 parameters, the independent ones. This applies similarly to individual column effect coefficients, etc.

The function requires an input which is the output of clustord, which includes the component out_parvec, the final vector of independent parameter values from the EM algorithm, which will correspond to a subset of the parameter values in out_parlist.

Functions

• calc_SE_rowcluster(): SE for rowclustering

• calc_SE_bicluster(): SE for biclustering