/// <summary>非線形最小二乗法テスト</summary>
public void Test()
{
Multifit mFit = new Multifit(Multifit.SolverType.LevenbergMarquardtWithScaling, 40, 3);
//学習データ作成
double[] ops = new double[40];
Random rnd = new Random();
//出力式
GSLFunctions.DArrayToDArray fnc = delegate(double[] inputs, ref double[] outputs)
{
double a = inputs[0];
double lambda = inputs[1];
double b = inputs[2];
for (int i = 0; i < outputs.Length; i++)
{
double t = i;
double yi = a * Math.Exp(-lambda * t) + b;
outputs[i] = yi - (1.0 + 5 * Math.Exp(-0.1 * t));
}
return GSLNET.ErrorNumber.GSL_SUCCESS;
};
int iterNum;
double[] param = new double[] { 1, 0, 0 };
GSLNET.ErrorNumber eNum = mFit.Solve(fnc, ref param, 1e-5, 500, out iterNum);
//メモリ解放
mFit.Dispose();
Console.WriteLine("終了");
Console.Read();
}
/// <summary>非線形最小二乗法テスト</summary>
public void Test()
{
Multifit mFit = new Multifit(Multifit.SolverType.LevenbergMarquardtWithScaling, 40, 3);
//学習データ作成
double[] ops = new double[40];
Random rnd = new Random();
//出力式
GSLFunctions.DArrayToDArray fnc = delegate(double[] inputs, ref double[] outputs)
{
double a = inputs[0];
double lambda = inputs[1];
double b = inputs[2];
for (int i = 0; i < outputs.Length; i++)
{
double t = i;
double yi = a * Math.Exp(-lambda * t) + b;
outputs[i] = yi - (1.0 + 5 * Math.Exp(-0.1 * t));
}
return(GSLNET.ErrorNumber.GSL_SUCCESS);
};
int iterNum;
double[] param = new double[] { 1, 0, 0 };
GSLNET.ErrorNumber eNum = mFit.Solve(fnc, ref param, 1e-5, 500, out iterNum);
//メモリ解放
mFit.Dispose();
Console.WriteLine("終了");
Console.Read();
}
