Skip to contents

Returns the lfsr for each single effect and outcome.

Usage

mvsusie_single_effect_lfsr(clfsr, alpha, lbf_outcome = NULL, min_lbf = NULL)

Arguments

clfsr

L x J x R conditional lfsr array.

alpha

L x J matrix of posterior inclusion probabilities.

lbf_outcome

L x R matrix of per-outcome conditional log BFs (from compute_per_outcome_lbf). If NULL (default), no filtering is applied.

min_lbf

Minimum per-outcome log BF threshold. Effects with lbf_outcome < min_lbf get lfsr set to 1. Default is NULL (no filtering). Set to 0 to filter when BF < 1 (data favors the null).

Value

L x R matrix of lfsr.

Details

When lbf_outcome is provided, outcomes where the per-outcome conditional log Bayes factor falls below min_lbf are set to lfsr = 1 (no evidence). This filters outcomes whose low lfsr comes purely from cross-outcome borrowing rather than per-outcome signal.