Fig. 1
From: Topic model-based mass spectrometric data analysis in cancer biomarker discovery studies
Graphical representation of the generative probabilistic model. Hyperparameters η, κ ′ together with sources of contaminants {β m } determine an average cancer profile γ ′. Each of the D profiles is associated with a mixture proportion θ d (regularized by hyperparameter α) and a topic panel consisting of {β m } and γ ′ (generated from the average cancer profile). Each of the N ions in a profile t n,d is sampled from a topic indicated by z n,d