Custom GLM Object: Spot Poisson — spot

Defines a Poisson-based GLM family object tailored for modeling spatial transcriptomics count data using deconvolution. Includes link functions, marginal mean, likelihood, gradient computation, and prediction utilities.

Usage

spot_poisson

Format

A list of class-specific functions:

link: Link and inverse link functions with derivatives.
marginal_mu: Marginal mean function.
original_class: Base GLM family used (Poisson).
predict: Predict means from X, beta, and lambda.
likelihood: Negative log-likelihood function.
grad: Gradient function with respect to beta.