program.cs

股票预测的程序

using System;
using System.Collections.Generic;
using System.Text;

namespace NeuralNetCS
{
    class Program
    {
        static double[] GetExamples()
        {
	        double[] examples = new double[101];
	        for( int x = 0; x < examples.Length; ++x )
	        {
		        double input = x / 100.0f;
		        if( input > 0.20 && input < 0.25 ||
			        input > 0.50 && input < 0.55 ||
			        input > 0.75 && input < 0.80 )
		        {
			        // This simulates noise in the input data
			        examples[x] = 0.3;
		        }
		        else
		        {
			        // The example function
			        examples[x] = 0.1 + (300 * Math.Sin((100 * input * input) / 3) + 95 * 
				        Math.Cos(100 * input) + (100 * input - 40) * (100 * input - 40) + 5) / 4000.0; 
		        }
	        }
	        return examples;
        }

        static bool HiddenNodesRangeFunc(int value)
        {
            return value > 0 && value <= 40;
        }

        static void Main()
        {
            Console.WriteLine("Artificial Intelligence: Neural Networks in C#");
            Console.WriteLine("Sven Groot");
            Console.WriteLine();
            int nodes = Util.InputValue<int>("Give the number of hidden nodes (1-40) (<enter> = 3): ", new Util.RangeFunc<int>(HiddenNodesRangeFunc), "The value must be between 1 and 40.", 3);
            uint epochs = Util.InputValue<uint>("Give the number of epochs (<enter> = 100000): ", new Util.RangeFunc<uint>(Util.Positive<uint>), "The value must be larger than 0.", 100000);
            double learningRate = Util.InputValue<double>("Give the learning rate (<enter> = 0.9): ", new Util.RangeFunc<double>(Util.Positive<double>), "The value must be larger than 0.", 0.9);
            Console.Write("Give the output file name (<enter> = 'net.out'): ");
            string fileName = Console.ReadLine().Trim();
            if( fileName.Length == 0 )
                fileName = "net.out";

            Net net = new Net(nodes, learningRate, GetExamples());
            net.Learn(epochs, fileName);
            Console.WriteLine();
            Console.WriteLine("Done!");
        }
    }
}