public void Test_LinearRegression()
{
//----- slope
{
double[,] testVectors =
{
{ 100.00000, 0.00000, 0.00000, 0.00000, 0.00000, },
{ 105.00000, 1.50000, 0.00000, 0.00000, 0.00000, },
{ 102.50000, 1.25000, 1.50000, 0.00000, 0.00000, },
{ 107.50000, 2.00000, 1.25000, 1.50000, 0.00000, },
{ 105.00000, 0.50000, 2.00000, 1.25000, 1.50000, },
{ 110.00000, 2.00000, 0.50000, 2.00000, 1.25000, },
{ 107.50000, 0.50000, 2.00000, 0.50000, 2.00000, },
{ 112.50000, 2.00000, 0.50000, 2.00000, 0.50000, },
{ 110.00000, 0.50000, 2.00000, 0.50000, 2.00000, },
{ 115.00000, 2.00000, 0.50000, 2.00000, 0.50000, },
};
TimeSeries <double> stimulus = new TimeSeries <double>();
string msg = "";
for (int row = 0; row <= testVectors.GetUpperBound(0); row++)
{
stimulus.Value = testVectors[row, 0];
ITimeSeries <double> response = stimulus.LinRegression(4).Slope;
msg += string.Format("{{{0:F5}, ", testVectors[row, 0]);
for (int col = 1; col <= testVectors.GetUpperBound(1); col++)
{
int t = col - 1;
double responseValue = response[t];
double expectedValue = testVectors[row, col];
msg += string.Format("{0:F5}, ", responseValue);
Assert.IsTrue(Math.Abs(responseValue - expectedValue) < 1e-5);
}
msg += "},\n";
}
Output.WriteLine(msg);
}
//----- intercept
{
double[,] testVectors =
{
{ 100.00000, 100.00000, 100.00000, 100.00000, 100.00000, },
{ 105.00000, 103.50000, 100.00000, 100.00000, 100.00000, },
{ 102.50000, 103.75000, 103.50000, 100.00000, 100.00000, },
{ 107.50000, 106.75000, 103.75000, 103.50000, 100.00000, },
{ 105.00000, 105.75000, 106.75000, 103.75000, 103.50000, },
{ 110.00000, 109.25000, 105.75000, 106.75000, 103.75000, },
{ 107.50000, 108.25000, 109.25000, 105.75000, 106.75000, },
{ 112.50000, 111.75000, 108.25000, 109.25000, 105.75000, },
{ 110.00000, 110.75000, 111.75000, 108.25000, 109.25000, },
{ 115.00000, 114.25000, 110.75000, 111.75000, 108.25000, },
};
TimeSeries <double> stimulus = new TimeSeries <double>();
string msg = "";
for (int row = 0; row <= testVectors.GetUpperBound(0); row++)
{
stimulus.Value = testVectors[row, 0];
ITimeSeries <double> response = stimulus.LinRegression(4).Intercept;
msg += string.Format("{{{0:F5}, ", testVectors[row, 0]);
for (int col = 1; col <= testVectors.GetUpperBound(1); col++)
{
int t = col - 1;
double responseValue = response[t];
double expectedValue = testVectors[row, col];
msg += string.Format("{0:F5}, ", responseValue);
Assert.IsTrue(Math.Abs(responseValue - expectedValue) < 1e-5);
}
msg += "},\n";
}
Output.WriteLine(msg);
}
//----- R2
{
double[,] testVectors =
{
{ 100.00000, 0.00000, 0.00000, 0.00000, 0.00000, },
{ 105.00000, 0.60000, 0.00000, 0.00000, 0.00000, },
{ 102.50000, 0.45455, 0.60000, 0.00000, 0.00000, },
{ 107.50000, 0.64000, 0.45455, 0.60000, 0.00000, },
{ 105.00000, 0.10000, 0.64000, 0.45455, 0.60000, },
{ 110.00000, 0.64000, 0.10000, 0.64000, 0.45455, },
{ 107.50000, 0.10000, 0.64000, 0.10000, 0.64000, },
{ 112.50000, 0.64000, 0.10000, 0.64000, 0.10000, },
{ 110.00000, 0.10000, 0.64000, 0.10000, 0.64000, },
{ 115.00000, 0.64000, 0.10000, 0.64000, 0.10000, },
};
TimeSeries <double> stimulus = new TimeSeries <double>();
string msg = "";
for (int row = 0; row <= testVectors.GetUpperBound(0); row++)
{
stimulus.Value = testVectors[row, 0];
ITimeSeries <double> response = stimulus.LinRegression(4).R2;
msg += string.Format("{{{0:F5}, ", testVectors[row, 0]);
for (int col = 1; col <= testVectors.GetUpperBound(1); col++)
{
int t = col - 1;
double responseValue = response[t];
double expectedValue = testVectors[row, col];
msg += string.Format("{0:F5}, ", responseValue);
Assert.IsTrue(Math.Abs(responseValue - expectedValue) < 1e-5);
}
msg += "},\n";
}
//Output.WriteLine(msg);
}
}
public void Test_LinearRegression()
{
//----- slope
{
double[,] testVectors =
{
{ 100.00000, 0.00000, 0.00000, 0.00000, 0.00000, },
{ 102.50000, 0.75000, 0.00000, 0.00000, 0.00000, },
{ 105.00000, 1.75000, 0.75000, 0.00000, 0.00000, },
{ 107.50000, 2.50000, 1.75000, 0.75000, 0.00000, },
{ 110.00000, 2.50000, 2.50000, 1.75000, 0.75000, },
{ 112.50000, 2.50000, 2.50000, 2.50000, 1.75000, },
{ 115.00000, 2.50000, 2.50000, 2.50000, 2.50000, },
{ 117.50000, 2.50000, 2.50000, 2.50000, 2.50000, },
{ 120.00000, 2.50000, 2.50000, 2.50000, 2.50000, },
{ 122.50000, 2.50000, 2.50000, 2.50000, 2.50000, },
};
TimeSeries <double> stimulus = new TimeSeries <double>();
for (int row = 0; row <= testVectors.GetUpperBound(0); row++)
{
stimulus.Value = testVectors[row, 0];
ITimeSeries <double> response = stimulus.LinRegression(4).Slope;
//Output.Write("{{{0:F5}, ", testVectors[row, 0]);
for (int col = 1; col <= testVectors.GetUpperBound(1); col++)
{
int t = col - 1;
double responseValue = response[t];
double expectedValue = testVectors[row, col];
//Output.Write("{0:F5}, ", responseValue);
Assert.IsTrue(Math.Abs(responseValue - expectedValue) < 1e-5);
}
//Output.WriteLine("},");
}
}
//----- intercept
{
double[,] testVectors =
{
{ 100.00000, 100.00000, 100.00000, 100.00000, 100.00000, },
{ 102.50000, 101.75000, 100.00000, 100.00000, 100.00000, },
{ 105.00000, 104.50000, 101.75000, 100.00000, 100.00000, },
{ 107.50000, 107.50000, 104.50000, 101.75000, 100.00000, },
{ 110.00000, 110.00000, 107.50000, 104.50000, 101.75000, },
{ 112.50000, 112.50000, 110.00000, 107.50000, 104.50000, },
{ 115.00000, 115.00000, 112.50000, 110.00000, 107.50000, },
{ 117.50000, 117.50000, 115.00000, 112.50000, 110.00000, },
{ 120.00000, 120.00000, 117.50000, 115.00000, 112.50000, },
{ 122.50000, 122.50000, 120.00000, 117.50000, 115.00000, },
};
TimeSeries <double> stimulus = new TimeSeries <double>();
for (int row = 0; row <= testVectors.GetUpperBound(0); row++)
{
stimulus.Value = testVectors[row, 0];
ITimeSeries <double> response = stimulus.LinRegression(4).Intercept;
//Output.Write("{{{0:F5}, ", testVectors[row, 0]);
for (int col = 1; col <= testVectors.GetUpperBound(1); col++)
{
int t = col - 1;
double responseValue = response[t];
double expectedValue = testVectors[row, col];
//Output.Write("{0:F5}, ", responseValue);
Assert.IsTrue(Math.Abs(responseValue - expectedValue) < 1e-5);
}
//Output.WriteLine("},");
}
}
//----- R2
{
double[,] testVectors =
{
{ 100.00000, 0.00000, 0.00000, 0.00000, 0.00000, },
{ 102.50000, 0.33333, 0.00000, 0.00000, 0.00000, },
{ 105.00000, 0.87755, 0.33333, 0.00000, 0.00000, },
{ 107.50000, 1.00000, 0.87755, 0.33333, 0.00000, },
{ 110.00000, 1.00000, 1.00000, 0.87755, 0.33333, },
{ 112.50000, 1.00000, 1.00000, 1.00000, 0.87755, },
{ 115.00000, 1.00000, 1.00000, 1.00000, 1.00000, },
{ 117.50000, 1.00000, 1.00000, 1.00000, 1.00000, },
{ 120.00000, 1.00000, 1.00000, 1.00000, 1.00000, },
{ 122.50000, 1.00000, 1.00000, 1.00000, 1.00000, },
};
TimeSeries <double> stimulus = new TimeSeries <double>();
for (int row = 0; row <= testVectors.GetUpperBound(0); row++)
{
stimulus.Value = testVectors[row, 0];
ITimeSeries <double> response = stimulus.LinRegression(4).R2;
//Output.Write("{{{0:F5}, ", testVectors[row, 0]);
for (int col = 1; col <= testVectors.GetUpperBound(1); col++)
{
int t = col - 1;
double responseValue = response[t];
double expectedValue = testVectors[row, col];
//Output.Write("{0:F5}, ", responseValue);
Assert.IsTrue(Math.Abs(responseValue - expectedValue) < 1e-5);
}
//Output.WriteLine("},");
}
}
}