Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!
Get Started for Free

layer_lines

Plot connection weights

Description

Plot connection weights in plotnet

Usage

layer_lines(mod_in, h_layer, layer1 = 1, layer2 = 2, out_layer = FALSE,
  nid, rel_rsc, all_in, pos_col, neg_col, x_range, y_range, line_stag, x_names,
  layer_x, struct, max_sp, prune_col = NULL, prune_lty = "dashed", skip)

Arguments

mod_in

neural network model object
h_layer

numeric indicating which connections to plot for the layer
layer1

numeric indicating order of first layer (for multiple hiden layers)
layer2

numeric indicating order of second layer (for multiple hiden layers)
out_layer

logical indicating if the lines are for the output layer
nid

logical value indicating if neural interpretation diagram is plotted, default TRUE
rel_rsc

numeric indicating the scaling range for the width of connection weights in a neural interpretation diagram. Default is NULL for no rescaling.
all_in

chr string indicating names of input variables for which connections are plotted, default all
pos_col

chr string indicating color of positive connection weights, default 'black'
neg_col

chr string indicating color of negative connection weights, default 'grey'
x_range

numeric of x axis range for base plot
y_range

numeric of y axis range for base plot
line_stag

numeric value that specifies distance of connection weights from nodes
x_names

chr string for names of input variables
layer_x

numeric indicating locations of layers on x axis
struct

numeric vector for network structure
max_sp

logical indicating if space is maximized in plot
prune_col

chr string indicating color of pruned connections, otherwise not shown
prune_lty

line type for pruned connections, passed to segments
skip

logical to plot connections for skip layer
NeuralNetTools

Visualization and Analysis Tools for Neural Networks

v1.5.2
CC0
Authors
Marcus W. Beck [aut, cre]
Initial release
2018-07-26

We don't support your browser anymore

Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.