update_signals: Propagation of fitted values through the graph.

In robertness/signalgraph: Signal Processing with Graphs

Description

'update_signals' uses graph propagation to update the fitted values (the output.signal attribute) at each vertex in a signalgraph object. It relies on the graph propagation function 'update_vertices' in the lucy package, using calculate_signal as the callback. The 'calculate_signal' callback, when applied to a given vertex, calculates the linear combination of fitted values from the parent nodes, and then applies the activation function to calculate the 'output.signal' attribute. The linear combination is calculated in the function 'getLinearCombination'.

Usage

update_signals(g, verbose = FALSE)

calculate_signal(g, v)

getLinearCombination(wts, model.mat)

Arguments

g	a signalgraph object.
v	the vertex to be updated given the parents
wts	the weights of the incoming edges given the parents
model.mat	a matrix constructing of the "output.signal" attributes of the parents

Value

a signal graph object with updated values for the signal attributes.

robertness/signalgraph documentation built on May 27, 2019, 10:33 a.m.