Module: Neuronet::NetworkStats

Included in:
Deep, FeedForward, MLP, Perceptron
Defined in:
lib/neuronet/network_stats.rb

Overview

Network Stats

Instance Method Summary collapse

Instance Method Details

#expected_nju ⇒ Object 



20
21
22
 
# File 'lib/neuronet/network_stats.rb', line 20

def expected_nju
  @expected_nju || expected_nju!
end

#expected_nju! ⇒ Object

See github.com/carlosjhr64/neuronet/wiki |𝝂| = π”ͺ + ΒΌβˆšπ‘*π”ͺβ€˜ + ΒΌβˆšπ‘*ΒΌβˆšπ‘β€™*π”ͺβ€œ + …



8
9
10
11
12
13
14
15
16
17
18
 
# File 'lib/neuronet/network_stats.rb', line 8

def expected_nju!
  nju = 0.0
  mult = 1.0
  reverse[1..].each do |layer|
    size = layer.size
    mju = 1 + (0.5 * size)
    nju += mult * mju
    mult *= 0.25 * Math.sqrt(size)
  end
  @expected_nju = nju
end

#njus ⇒ Object 



24
25
26
 
# File 'lib/neuronet/network_stats.rb', line 24

def njus
  output_layer.map(&:nju)
end