/// <summary>
/// Uses least squares linear regression to calculate the best fit sine wave for the given data.
/// </summary>
/// <param name="yValues">The y values of the data points.</param>
/// <param name="tValues">The time values of the data points, in seconds.</param>
/// <param name="frequency">The frequency of the sine wave, in Hz.</param>
/// <returns>A <see cref="SineWave"/> approximated from the given data points.</returns>
public static SineWave SineFit(double[] yValues, double[] tValues, double frequency)
{
double[] z = yValues;
double[] x = new double[tValues.Length];
double[] y = new double[tValues.Length];
double a, b, d;
double rad = 2.0D * Math.PI * frequency;
for (int i = 0; i < tValues.Length; i++)
{
double angle = rad * tValues[i];
x[i] = Math.Sin(angle);
y[i] = Math.Cos(angle);
}
CurveFit.LeastSquares(z, x, y, out d, out a, out b);
return(new SineWave
{
Amplitude = Math.Sqrt(a * a + b * b),
Frequency = frequency,
Phase = Math.Atan2(b, a),
Bias = d
});
}
/// <summary>
/// Uses least squares linear regression to calculate the best fit sine wave for the given data.
/// </summary>
/// <param name="yValues">The y values of the data points.</param>
/// <param name="tValues">The time values of the data points, in seconds.</param>
/// <param name="frequency">The frequency of the sine wave, in Hz.</param>
/// <returns>A <see cref="SineWave"/> approximated from the given data points.</returns>
public static SineWave SineFit(double[] yValues, double[] tValues, double frequency)
{
double[] z = yValues;
double[] x = tValues.Select(t => Math.Sin(2.0 * Math.PI * frequency * t)).ToArray();
double[] y = tValues.Select(t => Math.Cos(2.0 * Math.PI * frequency * t)).ToArray();
double a, b, d;
CurveFit.LeastSquares(z, x, y, out d, out a, out b);
return(new SineWave()
{
Amplitude = Math.Sqrt(a * a + b * b),
Frequency = frequency,
Phase = Math.Atan2(b, a),
Bias = d
});
}
private void PerformCalculation(object state)
{
try
{
m_calculating = true;
double[] xValues;
double[] yValues;
// Calculations are made against a copy of the current data set to keep lock time on
// data values down to a minimum. This allows data to be added with minimal delay.
lock (m_xValues)
{
xValues = m_xValues.ToArray();
yValues = m_yValues.ToArray();
}
// Takes new values and calculates slope (curve fit for 1st order polynomial).
m_slope = CurveFit.Compute(1, xValues, yValues)[1];
}
catch (Exception ex)
{
if ((object)Status != null)
{
Status(this, new EventArgs <string>("CurveFit failed: " + ex.Message));
}
}
finally
{
m_calculating = false;
}
if (Math.Sign(m_slope) != Math.Sign(m_lastSlope))
{
m_slopeRun = DateTime.UtcNow;
}
m_lastSlope = m_slope;
// Notifies consumer of new calculated slope.
if ((object)Recalculated != null)
{
Recalculated(this, EventArgs.Empty);
}
}
