vamf_stan: Varying-Censoring Aware Matrix Factorization via Stan.

In willtownes/vamf: Varying-Censoring Aware Matrix Factorization

Description

Runs VAMF with many random initializations in parallel. Parallelism will only work if the number of cores is set as a global option. For example options(mc.cores=4). Discards any random restarts that resulted in Stan errors, usually due to numeric convergence failures.

Usage

vamf_stan(ss, svmult = rep.int(1, 4), output_samples = 100)

Arguments

ss	A list of sufficient statistics from init_ss.
svmult	Vector of multipliers to increase or decrease the sigma_v scale hyperparameter. Length of svmult determines number of random restarts to run.
output_samples	How many samples from the approximate posterior to estimate the posterior mean of each parameter.

Value

List of results from each random initialization excluding numerical convergence failures. Each element of the list is an object with structure described in vb_wrap.

willtownes/vamf documentation built on May 8, 2019, 9:31 a.m.