backpropagation_example.rb

#
# Neural Network main file
#

# require File.dirname(__FILE__) + '/../../lib/ai4r/neural_network/backpropagation'
require File.dirname(__FILE__) + '/training_patterns'
require File.dirname(__FILE__) + '/patterns_with_noise'
require File.dirname(__FILE__) + '/patterns_with_base_noise'
require "rubygems"
require "ai4r"
require 'benchmark' # module for measuring time execution

times = Benchmark.measure do

    srand 1 # setting the random seed

    # creating network with 256 input-neurons, 3-neurons and 0 hidden layers
    net = Ai4r::NeuralNetwork::Backpropagation.new([256, 3])

    tr_input = TRIANGLE.flatten.collect { |input| input.to_f / 5.0}
    sq_input = SQUARE.flatten.collect { |input| input.to_f / 5.0}
    cr_input = CROSS.flatten.collect { |input| input.to_f / 5.0}

    tr_with_noise = TRIANGLE_WITH_NOISE.flatten.collect { |input| input.to_f / 5.0}
    sq_with_noise = SQUARE_WITH_NOISE.flatten.collect { |input| input.to_f / 5.0}
    cr_with_noise = CROSS_WITH_NOISE.flatten.collect { |input| input.to_f / 5.0}

    tr_with_base_noise = TRIANGLE_WITH_BASE_NOISE.flatten.collect { |input| input.to_f / 5.0}
    sq_with_base_noise = SQUARE_WITH_BASE_NOISE.flatten.collect { |input| input.to_f / 5.0}
    cr_with_base_noise = CROSS_WITH_BASE_NOISE.flatten.collect { |input| input.to_f / 5.0}

    puts "Training the network, please wait."
    puts "\nERROR...........................TRIANGLE................SQUARE..................CROSS................."
    101.times do |i|
      error1 = net.train(tr_input, [1,0,0])
      error2 = net.train(sq_input, [0,1,0])
      error3 = net.train(cr_input, [0,0,1])
      puts "Error after iteration #{i}:\t#{error1}\t#{error2}\t#{error3}" if i%20 == 0
    end

    def result_label(result)
      if result[0] > result[1] && result[0] > result[2]
        "TRIANGLE"
      elsif result[1] > result[2]
        "SQUARE"
      else
        "CROSS"
      end
    end

    puts "\nTraining Examples"
    puts "#{net.eval(tr_input).inspect} => #{result_label(net.eval(tr_input))}"
    puts "#{net.eval(sq_input).inspect} => #{result_label(net.eval(sq_input))}"
    puts "#{net.eval(cr_input).inspect} => #{result_label(net.eval(cr_input))}"
    puts "Examples with noise"
    puts "#{net.eval(tr_with_noise).inspect} => #{result_label(net.eval(tr_with_noise))}"
    puts "#{net.eval(sq_with_noise).inspect} => #{result_label(net.eval(sq_with_noise))}"
    puts "#{net.eval(cr_with_noise).inspect} => #{result_label(net.eval(cr_with_noise))}"
    puts "Examples with base noise"
    puts "#{net.eval(tr_with_base_noise).inspect} => #{result_label(net.eval(tr_with_base_noise))}"
    puts "#{net.eval(sq_with_base_noise).inspect} => #{result_label(net.eval(sq_with_base_noise))}"
    puts "#{net.eval(cr_with_base_noise).inspect} => #{result_label(net.eval(cr_with_base_noise))}"

end

puts "\nElapsed time: #{times}"