Figure 1
From: NETGEM: Network Embedded Temporal GEnerative Model for gene expression data
The NETGEM model. This figure presents a high-level description of the NETGEM model and the resulting inference algorithm. Here Λ, Θ are hyper-parameters and β is the strain damping factor. The parameters to be learnt are the functional category component for each edge (α
e
) and the functional category transition probability matrix (Q
h
). The inference is done over the hidden variables 
, where 
and 
are the random variables corresponding to the active functional category and the interaction strength for edge e at time t respectively. The inferred interaction dynamics 
is used to compute the edge and functional category change scores s
T
(e) and 
respectively.