Example usage for org.apache.commons.math3.fitting.leastsquares EvaluationRmsChecker EvaluationRmsChecker

Introduction

In this page you can find the example usage for org.apache.commons.math3.fitting.leastsquares EvaluationRmsChecker EvaluationRmsChecker.

Prototype

public EvaluationRmsChecker(final double relTol, final double absTol) 

Source Link

Document

Create a convergence checker for the RMS with a relative and absolute tolerance.

Usage

From source file:uk.ac.diamond.scisoft.analysis.optimize.ApacheOptimizer.java

/**
 * create a multivariateJacobianFunction from MVF and MMF (using builder?)
 * //  ww  w. j  a  v  a  2s .c  o  m
 */
private void internalLeastSquaresOptimize() {
    LeastSquaresOptimizer opt = createLeastSquaresOptimizer();

    try {

        LeastSquaresBuilder builder = new LeastSquaresBuilder().model(createJacobianFunction())
                .target(data.getData()).start(getParameterValues()).lazyEvaluation(false)
                .maxEvaluations(MAX_EVAL).maxIterations(MAX_ITER);

        builder.checker(new EvaluationRmsChecker(REL_TOL, ABS_TOL));

        if (weight != null) {
            builder.weight(MatrixUtils.createRealDiagonalMatrix(weight.getData()));
        }

        // TODO add checker, validator
        LeastSquaresProblem problem = builder.build();

        Optimum result = opt.optimize(problem);

        RealVector res = result.getPoint();
        setParameterValues(
                res instanceof ArrayRealVector ? ((ArrayRealVector) res).getDataRef() : res.toArray());
        try {
            RealVector err = result.getSigma(1e-14);

            //            sqrt(S / (n - m) * C[i][i]);
            double c = result.getCost();
            int n = data.getSize();
            int m = getParameterValues().length;

            double[] s = err instanceof ArrayRealVector ? ((ArrayRealVector) err).getDataRef() : err.toArray();

            errors = new double[s.length];

            for (int i = 0; i < errors.length; i++)
                errors[i] = Math.sqrt(((c * c) / ((n - m)) * (s[i] * s[i])));

        } catch (SingularMatrixException e) {
            logger.warn("Could not find errors as covariance matrix was singular");
        }

        logger.trace("Residual: {} from {}", result.getRMS(), Math.sqrt(calculateResidual()));
    } catch (Exception e) {
        logger.error("Problem with least squares optimizer", e);
        throw new IllegalArgumentException("Problem with least squares optimizer");
    }
}

---