KickJava   Java API By Example, From Geeks To Geeks.

Java > Open Source Codes > org > jfree > data > statistics > Regression


1 /* ===========================================================
2  * JFreeChart : a free chart library for the Java(tm) platform
3  * ===========================================================
4  *
5  * (C) Copyright 2000-2005, by Object Refinery Limited and Contributors.
6  *
7  * Project Info: http://www.jfree.org/jfreechart/index.html
8  *
9  * This library is free software; you can redistribute it and/or modify it
10  * under the terms of the GNU Lesser General Public License as published by
11  * the Free Software Foundation; either version 2.1 of the License, or
12  * (at your option) any later version.
13  *
14  * This library is distributed in the hope that it will be useful, but
15  * WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
16  * or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
17  * License for more details.
18  *
19  * You should have received a copy of the GNU Lesser General Public
20  * License along with this library; if not, write to the Free Software
21  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301,
22  * USA.
23  *
24  * [Java is a trademark or registered trademark of Sun Microsystems, Inc.
25  * in the United States and other countries.]
26  *
27  * ---------------
28  * Regression.java
29  * ---------------
30  * (C) Copyright 2002-2005, by Object Refinery Limited.
31  *
32  * Original Author: David Gilbert (for Object Refinery Limited);
33  * Contributor(s): -;
34  *
35  * $Id: Regression.java,v 1.3.2.1 2005/10/25 21:34:46 mungady Exp $
36  *
37  * Changes
38  * -------
39  * 30-Sep-2002 : Version 1 (DG);
40  * 18-Aug-2003 : Added 'abstract' (DG);
41  * 15-Jul-2004 : Switched getX() with getXValue() and getY() with
42  * getYValue() (DG);
43  *
44  */
45
46 package org.jfree.data.statistics;
47
48 import org.jfree.data.xy.XYDataset;
49
50 /**
51  * A utility class for fitting regression curves to data.
52  */
53 public abstract class Regression {
54
55     /**
56      * Returns the parameters 'a' and 'b' for an equation y = a + bx, fitted to
57      * the data using ordinary least squares regression. The result is
58      * returned as a double[], where result[0] --> a, and result[1] --> b.
59      *
60      * @param data the data.
61      *
62      * @return The parameters.
63      */
64     public static double[] getOLSRegression(double[][] data) {
65
66         int n = data.length;
67         if (n < 2) {
68             throw new IllegalArgumentException JavaDoc("Not enough data.");
69         }
70
71         double sumX = 0;
72         double sumY = 0;
73         double sumXX = 0;
74         double sumXY = 0;
75         for (int i = 0; i < n; i++) {
76             double x = data[i][0];
77             double y = data[i][1];
78             sumX += x;
79             sumY += y;
80             double xx = x * x;
81             sumXX += xx;
82             double xy = x * y;
83             sumXY += xy;
84         }
85         double sxx = sumXX - (sumX * sumX) / n;
86         double sxy = sumXY - (sumX * sumY) / n;
87         double xbar = sumX / n;
88         double ybar = sumY / n;
89
90         double[] result = new double[2];
91         result[1] = sxy / sxx;
92         result[0] = ybar - result[1] * xbar;
93
94         return result;
95
96     }
97
98     /**
99      * Returns the parameters 'a' and 'b' for an equation y = a + bx, fitted to
100      * the data using ordinary least squares regression. The result is returned
101      * as a double[], where result[0] --> a, and result[1] --> b.
102      *
103      * @param data the data.
104      * @param series the series (zero-based index).
105      *
106      * @return The parameters.
107      */
108     public static double[] getOLSRegression(XYDataset data, int series) {
109
110         int n = data.getItemCount(series);
111         if (n < 2) {
112             throw new IllegalArgumentException JavaDoc("Not enough data.");
113         }
114
115         double sumX = 0;
116         double sumY = 0;
117         double sumXX = 0;
118         double sumXY = 0;
119         for (int i = 0; i < n; i++) {
120             double x = data.getXValue(series, i);
121             double y = data.getYValue(series, i);
122             sumX += x;
123             sumY += y;
124             double xx = x * x;
125             sumXX += xx;
126             double xy = x * y;
127             sumXY += xy;
128         }
129         double sxx = sumXX - (sumX * sumX) / n;
130         double sxy = sumXY - (sumX * sumY) / n;
131         double xbar = sumX / n;
132         double ybar = sumY / n;
133
134         double[] result = new double[2];
135         result[1] = sxy / sxx;
136         result[0] = ybar - result[1] * xbar;
137
138         return result;
139
140     }
141
142     /**
143      * Returns the parameters 'a' and 'b' for an equation y = ax^b, fitted to
144      * the data using a power regression equation. The result is returned as
145      * an array, where double[0] --> a, and double[1] --> b.
146      *
147      * @param data the data.
148      *
149      * @return The parameters.
150      */
151     public static double[] getPowerRegression(double[][] data) {
152
153         int n = data.length;
154         if (n < 2) {
155             throw new IllegalArgumentException JavaDoc("Not enough data.");
156         }
157
158         double sumX = 0;
159         double sumY = 0;
160         double sumXX = 0;
161         double sumXY = 0;
162         for (int i = 0; i < n; i++) {
163             double x = Math.log(data[i][0]);
164             double y = Math.log(data[i][1]);
165             sumX += x;
166             sumY += y;
167             double xx = x * x;
168             sumXX += xx;
169             double xy = x * y;
170             sumXY += xy;
171         }
172         double sxx = sumXX - (sumX * sumX) / n;
173         double sxy = sumXY - (sumX * sumY) / n;
174         double xbar = sumX / n;
175         double ybar = sumY / n;
176
177         double[] result = new double[2];
178         result[1] = sxy / sxx;
179         result[0] = Math.pow(Math.exp(1.0), ybar - result[1] * xbar);
180
181         return result;
182
183     }
184
185     /**
186      * Returns the parameters 'a' and 'b' for an equation y = ax^b, fitted to
187      * the data using a power regression equation. The result is returned as
188      * an array, where double[0] --> a, and double[1] --> b.
189      *
190      * @param data the data.
191      * @param series the series to fit the regression line against.
192      *
193      * @return The parameters.
194      */
195     public static double[] getPowerRegression(XYDataset data, int series) {
196
197         int n = data.getItemCount(series);
198         if (n < 2) {
199             throw new IllegalArgumentException JavaDoc("Not enough data.");
200         }
201
202         double sumX = 0;
203         double sumY = 0;
204         double sumXX = 0;
205         double sumXY = 0;
206         for (int i = 0; i < n; i++) {
207             double x = Math.log(data.getXValue(series, i));
208             double y = Math.log(data.getYValue(series, i));
209             sumX += x;
210             sumY += y;
211             double xx = x * x;
212             sumXX += xx;
213             double xy = x * y;
214             sumXY += xy;
215         }
216         double sxx = sumXX - (sumX * sumX) / n;
217         double sxy = sumXY - (sumX * sumY) / n;
218         double xbar = sumX / n;
219         double ybar = sumY / n;
220
221         double[] result = new double[2];
222         result[1] = sxy / sxx;
223         result[0] = Math.pow(Math.exp(1.0), ybar - result[1] * xbar);
224
225         return result;
226
227     }
228
229 }
230
Popular Tags