コード例 #1
0
        private void button2_Click(object sender, EventArgs e)
        {
            double lerningSpeed = 0.1;
            bool   osiagnieto1 = true, osiagnieto2 = true, osiagnieto3 = true;
            /////WCZYTYWANIE DANYCH//////
            string        path       = @"1.txt";
            List <double> inputList  = new List <double>();;
            List <double> outputList = new List <double>();

            ReadFile(@"1.txt", inputList, outputList);
            List <double> testNumbers = new List <double>();
            List <double> testOutputs = new List <double>();

            ReadFile(@"test.txt", testNumbers, testOutputs);
            /////////////////////////////
            int liczba_danych1 = inputList.Count;

            double[] inputData = new double[liczba_danych1];
            inputList.CopyTo(inputData);
            int inputAmount  = 1;
            int outputAmount = 1;

            double[] outputData = new double[liczba_danych1];
            outputList.CopyTo(outputData);
            IFunc  inputFunction   = new Linear();
            IFunc  sigmoidFunction = new Sigmoid();
            IFunc  outputFunction  = new Linear();
            double error           = 0;
            MLP    s1 = new MLP(inputAmount, getNeuralAmount(), outputAmount, lerningSpeed, inputFunction, sigmoidFunction, outputFunction, Bias.Checked);


            Wykres.Series["Dane 1"].Points.Clear();
            Wykres.Series["Dane 2"].Points.Clear();
            Wykres.Series["Dane 3"].Points.Clear();
            for (int i = 0; i < 10000; i++)
            {
                s1.MultiLayerPerceptron(1, inputData, outputData);
                error = s1.GetGlobalError();

                //               if (error > 0.001)
                //                   break;
            }
            //DANE TESTWOE

            double[] datain = new double[testNumbers.Count];
            testNumbers.CopyTo(datain);
            double[] dataout = new double[testNumbers.Count];
            testOutputs.CopyTo(dataout);
            for (int i = 0; i < testNumbers.Count; i++)
            {
                Wykres.Series["Dane 1"].Points.AddXY(datain[i], dataout[i]);
            }

            for (int i = 0; i < testNumbers.Count; i++)
            {
                Wykres.Series["Dane 2"].Points.AddXY(datain[i], s1.CountOne(datain[i]));
            }
        }
コード例 #2
0
        //ZAD 1
        private void button1_Click(object sender, EventArgs e)
        {
            double lerningSpeed = 0.3;
            //Wykresy
            double lerningSpeed2 = 0.3;
            double lerningSpeed3 = 0.3;
            double error2 = 0;
            double error3 = 0;
            bool   osiagnieto1 = true, osiagnieto2 = true, osiagnieto3 = true;

            /////////////////////////////
            double[,] inputData = new double[4, 4]
            {
                { 1, 0, 0, 0 }, { 0, 1, 0, 0 }, { 0, 0, 1, 0 }, { 0, 0, 0, 1 }
            };
            double[][] testy = new double[4][];
            for (int i = 0; i < 4; i++)
            {
                testy[i] = new double[4] {
                    0, 0, 0, 0
                };
                testy[i][i] = 1;
            }

            int inputAmount  = 4;
            int outputAmount = 4;

            double[,] outputData = inputData;
            IFunc  inputFunction   = new Linear();
            IFunc  sigmoidFunction = new Sigmoid();
            IFunc  outputFunction  = new Sigmoid();
            double error           = 0;
            MLP    s1 = new MLP(inputAmount, getNeuralAmount(), outputAmount, lerningSpeed, inputFunction, sigmoidFunction, outputFunction, Bias.Checked);

            //Wykresy
            //MLP s1 = new MLP(inputAmount, 1, outputAmount, lerningSpeed, inputFunction, sigmoidFunction, outputFunction, Bias.Checked);
            // MLP s2 = new MLP(inputAmount, getNeuralAmount(), outputAmount, lerningSpeed2, inputFunction, sigmoidFunction, outputFunction,  Bias.Checked);
            // MLP s3 = new MLP(inputAmount, getNeuralAmount(), outputAmount, lerningSpeed3, inputFunction, sigmoidFunction, outputFunction,  Bias.Checked);
            //////////////////////////
            //           Wykres.Series.Add("Lerning Rate 0.3");
            //           Wykres.Series.Add("Lerning Rate 0.6");
            //           Wykres.Series.Add("Lerning Rate 0.9");

            /*
             * Wykres.Series["Lerning Rate 0.3"].Points.Clear();
             * Wykres.Series["Lerning Rate 0.6"].Points.Clear();
             * Wykres.Series["Lerning Rate 0.9"].Points.Clear();
             */

            Wykres.Series["Dane 1"].Points.Clear();
            Wykres.Series["Dane 2"].Points.Clear();
            Wykres.Series["Dane 3"].Points.Clear();

            for (int i = 0; i < 10000; i++)
            {
                s1.MultiLayerPerceptron(4, inputData, outputData);
                error = s1.GetGlobalError();
                Wykres.Series["Dane 1"].Points.AddXY(i, error);

                /*//  s2.MultiLayerPerceptron(4, inputData, outputData);
                 * error2 = s2.GetGlobalError();
                 * Wykres.Series["Dane 2"].Points.AddXY(i, error2);
                 * s3.MultiLayerPerceptron(4, inputData, outputData);
                 * error3 = s3.GetGlobalError();
                 * Wykres.Series["Dane 3"].Points.AddXY(i, error3);
                 * /////////*/

                // Do Labeli
                if (error <= 0.01 && osiagnieto1)
                {
                    label22.Text = i.ToString(); osiagnieto1 = false;
                }
                if (error2 <= 0.01 && osiagnieto2)
                {
                    label23.Text = i.ToString(); osiagnieto2 = false;
                }
                if (error3 <= 0.01 && osiagnieto3)
                {
                    label24.Text = i.ToString(); osiagnieto3 = false;
                }
            }
            List <double> teemp = s1.getWeight();

            saveFile(@"wagi.txt", teemp);
            teemp.Clear();
            for (int i = 0; i < 4; i++)
            {
                double[] temp = s1.Count(testy[i]);
                foreach (double x in temp)
                {
                    teemp.Add(x);
                }
            }
            saveFile(@"wyniki.txt", teemp);
            double[] input   = new double[4];
            double[] results = new double[16];
            for (int j = 0; j < 4; j++)
            {
                for (int i = 0; i < 4; i++)
                {
                    input[i] = inputData[i, j];
                }
                double[] output = new double[4];
                output = s1.Count2(input);
                for (int i = 0; i < getNeuralAmount(); i++)
                {
                    results[j * getNeuralAmount() + i] = output[i];
                }
                label1.Text  = String.Format("{0:N3}", results[0]);
                label2.Text  = String.Format("{0:N3}", results[1]);
                label3.Text  = String.Format("{0:N3}", results[2]);
                label4.Text  = String.Format("{0:N3}", results[3]);
                label8.Text  = String.Format("{0:N3}", results[4]);
                label7.Text  = String.Format("{0:N3}", results[5]);
                label6.Text  = String.Format("{0:N3}", results[6]);
                label5.Text  = String.Format("{0:N3}", results[7]);
                label9.Text  = String.Format("{0:N3}", results[8]);
                label10.Text = String.Format("{0:N3}", results[9]);
                label11.Text = String.Format("{0:N3}", results[10]);
                label12.Text = String.Format("{0:N3}", results[11]);
            }
        }