Return the estimated mixture proportions
get_estimated_pi(m, dimension = c("cov", "grid", "all"))the mash result
indicates whether you want the mixture proportions for the covariances, grid, or all
a named vector containing the estimated mixture proportions.
If the fit was done with `usepointmass=TRUE` then the first element of the returned vector will correspond to the null, and the remaining elements to the non-null covariance matrices. Suppose the fit was done with $K$ covariances and a grid of length $L$. If `dimension=cov` then the returned vector will be of length $K$ (or $K+1$ if `usepointmass=TRUE`). If `dimension=grid` then the returned vector will be of length $L$ (or $L+1$). If `dimension=all` then the returned vector will be of length $LK$ (or $LK+1$). The names of the vector will be informative for which combination each element corresponds to.
simdata = simple_sims(50,5,1)
data = mash_set_data(simdata$Bhat, simdata$Shat)
m = mash(data, cov_canonical(data))
#> - Computing 200 x 141 likelihood matrix.
#> - Likelihood calculations took 0.01 seconds.
#> - Fitting model with 141 mixture components.
#> - Model fitting took 0.15 seconds.
#> - Computing posterior matrices.
#> - Computation allocated took 0.00 seconds.
get_estimated_pi(m)
#> null identity condition_1 condition_2 condition_3
#> 0.59193141 0.04017492 0.03668422 0.00000000 0.00000000
#> condition_4 condition_5 equal_effects simple_het_1 simple_het_2
#> 0.00000000 0.03834005 0.04235686 0.16733138 0.08318115
#> simple_het_3
#> 0.00000000