KronFastScanner¶

class glimix_core.lmm.KronFastScanner(Y, A, X, G, terms)[source]¶

Approximated fast inference over several covariates.

Specifically, it maximizes the log of the marginal likelihood

log(p(Y)ⱼ) = log𝓝(vec(Y) | (A ⊗ X)vec(𝚩ⱼ) + (Aⱼ ⊗ Xⱼ)vec(𝚨ⱼ), sⱼK),

where K = C₀ ⊗ GGᵀ + C₁ ⊗ I and ⱼ index the candidates set. For performance purpose, we optimise only the fixed-effect sizes and scale parameters. Therefore, K is fixed throughout the process.

__init__(Y, A, X, G, terms)[source]¶

Constructor.

Parameters

Y ((n, p) array_like) – Outcome matrix.
A ((n, n) array_like) – Trait-by-trait design matrix.
X ((n, c) array_like) – Covariates design matrix.
G ((n, r) array_like) – Matrix G from the GGᵀ term.
terms (dict) – Pre-computed terms.

Methods

`__init__`(Y, A, X, G, terms)	Constructor.
`null_lml`()
`scan`(A1, X1)	LML, fixed-effect sizes, and scale of the candidate set.

Attributes

`null_beta`	Optimal 𝛃 according to the marginal likelihood.
`null_beta_covariance`	Covariance of the optimal 𝛃 according to the marginal likelihood.
`null_beta_se`	Standard errors of the optimal 𝛃.
`null_scale`	Optimal s according to the marginal likelihood.