A variational Bayesian algorithm, based on the Laplace Spike-and-Slab prior, is tailored for multi-source heterogeneous models and focuses on variable selection exclusively for the homogeneous covariates.
vb_lap_local(X, Z, Y, max_iter = 1000, tol = 1e-06, a = 1, b = 10, lambda = 1)The mean of the homogeneity coefficient:mu; The variance of homogeneity coefficient:sigma; Selection coefficient:gamma; Mean and covariance of heterogeneity coefficients:m, s2.
Homogeneous covariates
Heterogeneous covariates
Response covariates
Maximum number of iterations, Defaut:1000
Algorithm convergence tolerance, Defaut:1e-6
A prior of Beta distribution, Defaut:1
A prior of Beta distribution, Defaut:10
A prior of Laplace distribution, Defaut:1