EncogNeuralNetworkSumNumbers.cs

using Encog;
using Encog.ML.Data;
using Encog.ML.Data.Basic;
using Encog.ML.Train;
using Encog.Neural.Networks;
using Encog.Neural.Networks.Layers;
using Encog.Neural.Networks.Training.Propagation.Resilient;
using System;
using System.Collections.Generic;
using Encog.Engine.Network.Activation;

namespace EncogDotNetPlay
{
    internal class Program
    {
        static Func<double, double, double> sum = (a, b) => a + b;
        static Func<double, double, double> mult = (a, b) => a * b;

        private static void Main()
        {
            var rand = new Random();
            var op = sum;
            var trainingData = new List<IMLDataPair>();
            for (var i = 0; i < 1000; i++)
            {
                var first = rand.Next(1, 100);
                var second = rand.Next(1, 100);
                var input = new double[] { first, second };
                var output = new[] { op(first, second) };
                var mldatapair = new BasicMLDataPair(new BasicMLData(input), new BasicMLData(output));
                trainingData.Add(mldatapair);
            }

            // create a neural network, without using a factory
            var network = new BasicNetwork();
            for (var i = 0; i < 1; i++)
            {
                //network.AddLayer(new BasicLayer(new ActivationLOG(), true, 2));
                //network.AddLayer(new BasicLayer(new ActivationLOG(), false, 1));
                //network.AddLayer(new BasicLayer(new ActivationElliott(), true, 2));
                network.AddLayer(new BasicLayer(null, true, 2));
                network.AddLayer(new BasicLayer(null, false, 1));
            }
            network.Structure.FinalizeStructure();
            network.Reset();

            // train the neural network
            IMLTrain train = new ResilientPropagation(network, new BasicMLDataSet(trainingData));

            var epoch = 1;
            do
            {
                train.Iteration();
                Console.WriteLine(@"Epoch #" + epoch + @" Error:" + train.Error);
                epoch++;
            } while (train.Error > 0.0000001 && epoch < 10000);

            train.FinishTraining();

            // test the neural network
            Console.WriteLine(@"Neural Network Results:");
            foreach (var pair in trainingData)
            {
                var output = network.Compute(pair.Input);
                Console.WriteLine(pair.Input + @", actual=" + output + @", ideal=" + pair.Ideal[0]);
            }

            var data = new List<BasicMLData>
            {
                new BasicMLData(new[] {1.0, 2.0}),
                new BasicMLData(new[] {11.0, 21.0}),
                new BasicMLData(new[] {1.0, -20.0}),
                new BasicMLData(new[] {-20.0, -20.0}),
                new BasicMLData(new[] {-20.0, 20.0}),
                new BasicMLData(new[] {20.0, -20.0}),
                new BasicMLData(new[] {20.0, 20.0}),
                new BasicMLData(new[] {1337.0, 5015})
            };

            foreach (var item in data)
            {
                var output = network.Compute(item);
                Console.WriteLine(item + @", actual=" + output + $" {op(item[0], item[1])}");
            }

            EncogFramework.Instance.Shutdown();
        }
    }
}

xwvwxv