private DoubleRange computeInterval(double[] input, int numberOfSamples, double percent, double se)
{
double y = linear.Transform(input);
double df = GetDegreesOfFreedom(numberOfSamples);
var t = new TTest(estimatedValue: y, standardError: se, degreesOfFreedom: df);
DoubleRange lci = t.GetConfidenceInterval(percent);
DoubleRange nci = new DoubleRange(linkFunction.Inverse(lci.Min), linkFunction.Inverse(lci.Max));
return(nci);
}
public double Compute(double[] input)
{
double sum = coefficients[0];
for (int i = 1; i < coefficients.Length; i++)
{
sum += input[i - 1] * coefficients[i];
}
return(linkFunction.Inverse(sum));
}
/// <summary>
/// Computes the given input to produce the corresponding output.
/// </summary>
///
/// <remarks>
/// For a binary decision problem, the decision for the negative
/// or positive class is typically computed by taking the sign of
/// the machine's output.
/// </remarks>
///
/// <param name="inputs">An input vector.</param>
/// <param name="output">The output of the machine. If this is a
/// <see cref="IsProbabilistic">probabilistic</see> machine, the
/// output is the probability of the positive class. If this is
/// a standard machine, the output is the distance to the decision
/// hyperplane in feature space.</param>
///
/// <returns>The decision label for the given input.</returns>
///
public virtual int Compute(double[] inputs, out double output)
{
output = threshold;
if (supportVectors == null)
{
for (int i = 0; i < weights.Length; i++)
{
output += weights[i] * inputs[i];
}
}
else
{
for (int i = 0; i < supportVectors.Length; i++)
{
double sum = 0;
for (int j = 0; j < inputs.Length; j++)
{
sum += supportVectors[i][j] * inputs[j];
}
output += weights[i] * sum;
}
}
if (IsProbabilistic)
{
output = linkFunction.Inverse(output);
return(output >= 0.5 ? +1 : -1);
}
return(output >= 0 ? +1 : -1);
}
