001 /*
002 * Licensed to the Apache Software Foundation (ASF) under one or more
003 * contributor license agreements. See the NOTICE file distributed with
004 * this work for additional information regarding copyright ownership.
005 * The ASF licenses this file to You under the Apache License, Version 2.0
006 * (the "License"); you may not use this file except in compliance with
007 * the License. You may obtain a copy of the License at
008 *
009 * http://www.apache.org/licenses/LICENSE-2.0
010 *
011 * Unless required by applicable law or agreed to in writing, software
012 * distributed under the License is distributed on an "AS IS" BASIS,
013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
014 * See the License for the specific language governing permissions and
015 * limitations under the License.
016 */
017
018 package org.apache.commons.math.optimization.fitting;
019
020 import org.apache.commons.math.FunctionEvaluationException;
021 import org.apache.commons.math.analysis.polynomials.PolynomialFunction;
022 import org.apache.commons.math.optimization.DifferentiableMultivariateVectorialOptimizer;
023 import org.apache.commons.math.optimization.OptimizationException;
024
025 /** This class implements a curve fitting specialized for polynomials.
026 * <p>Polynomial fitting is a very simple case of curve fitting. The
027 * estimated coefficients are the polynomial coefficients. They are
028 * searched by a least square estimator.</p>
029 * @version $Revision: 1073270 $ $Date: 2011-02-22 10:19:27 +0100 (mar. 22 f??vr. 2011) $
030 * @since 2.0
031 */
032
033 public class PolynomialFitter {
034
035 /** Fitter for the coefficients. */
036 private final CurveFitter fitter;
037
038 /** Polynomial degree. */
039 private final int degree;
040
041 /** Simple constructor.
042 * <p>The polynomial fitter built this way are complete polynomials,
043 * ie. a n-degree polynomial has n+1 coefficients.</p>
044 * @param degree maximal degree of the polynomial
045 * @param optimizer optimizer to use for the fitting
046 */
047 public PolynomialFitter(int degree, final DifferentiableMultivariateVectorialOptimizer optimizer) {
048 this.fitter = new CurveFitter(optimizer);
049 this.degree = degree;
050 }
051
052 /** Add an observed weighted (x,y) point to the sample.
053 * @param weight weight of the observed point in the fit
054 * @param x abscissa of the point
055 * @param y observed value of the point at x, after fitting we should
056 * have P(x) as close as possible to this value
057 */
058 public void addObservedPoint(double weight, double x, double y) {
059 fitter.addObservedPoint(weight, x, y);
060 }
061
062 /**
063 * Remove all observations.
064 * @since 2.2
065 */
066 public void clearObservations() {
067 fitter.clearObservations();
068 }
069
070 /** Get the polynomial fitting the weighted (x, y) points.
071 * @return polynomial function best fitting the observed points
072 * @exception OptimizationException if the algorithm failed to converge
073 */
074 public PolynomialFunction fit() throws OptimizationException {
075 try {
076 return new PolynomialFunction(fitter.fit(new ParametricPolynomial(), new double[degree + 1]));
077 } catch (FunctionEvaluationException fee) {
078 // should never happen
079 throw new RuntimeException(fee);
080 }
081 }
082
083 /** Dedicated parametric polynomial class. */
084 private static class ParametricPolynomial implements ParametricRealFunction {
085
086 /** {@inheritDoc} */
087 public double[] gradient(double x, double[] parameters) {
088 final double[] gradient = new double[parameters.length];
089 double xn = 1.0;
090 for (int i = 0; i < parameters.length; ++i) {
091 gradient[i] = xn;
092 xn *= x;
093 }
094 return gradient;
095 }
096
097 /** {@inheritDoc} */
098 public double value(final double x, final double[] parameters) {
099 double y = 0;
100 for (int i = parameters.length - 1; i >= 0; --i) {
101 y = y * x + parameters[i];
102 }
103 return y;
104 }
105
106 }
107
108 }