JMSLTM Numerical Library 5.0.1

com.imsl.stat
Interface NonlinearRegression.Derivative

All Superinterfaces:
NonlinearRegression.Function
Enclosing class:
NonlinearRegression

public static interface NonlinearRegression.Derivative
extends NonlinearRegression.Function

Public interface for the user supplied function to compute the derivative for NonlinearRegression.


Method Summary
 boolean derivative(double[] theta, int iobs, double[] frq, double[] wt, double[] de)
          Computes the weight, frequency, and partial derivatives of the residual given the parameter vector theta for a single observation.
 
Methods inherited from interface com.imsl.stat.NonlinearRegression.Function
f
 

Method Detail

derivative

boolean derivative(double[] theta,
                   int iobs,
                   double[] frq,
                   double[] wt,
                   double[] de)
Computes the weight, frequency, and partial derivatives of the residual given the parameter vector theta for a single observation.

Parameters:
theta - An input double array which contains the parameter values of the regression function. The length of theta corresponds to the number of unknown parameters in the regression function.
iobs - An input int value indicating the observation index. The function is evaluated at observation y[iobs].
frq - An output double array of length 1 containing the frequency for observation y[iobs].
wt - An output double array of length 1 containing the weight for the observation y[iobs]. Use wt = 1.0 for equal weighting (unweighted least squares).
de - An output double array containing the partial derivatives of the error (residual) for observation y[iobs]. The length of de corresponds to the number of unknown parameters in the regression function.
Returns:
A boolean value representing the completion indicator. true indicates iobs is less than the number of observations. false indicates iobs is greater than or equal to the number of observations and wt, freq, and de are not output.

JMSLTM Numerical Library 5.0.1

Copyright © 1970-2008 Visual Numerics, Inc.
Built July 8 2008.