View Javadoc
1   /*
2    * Licensed to the Apache Software Foundation (ASF) under one or more
3    * contributor license agreements.  See the NOTICE file distributed with
4    * this work for additional information regarding copyright ownership.
5    * The ASF licenses this file to You under the Apache License, Version 2.0
6    * (the "License"); you may not use this file except in compliance with
7    * the License.  You may obtain a copy of the License at
8    *
9    *      https://www.apache.org/licenses/LICENSE-2.0
10   *
11   * Unless required by applicable law or agreed to in writing, software
12   * distributed under the License is distributed on an "AS IS" BASIS,
13   * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14   * See the License for the specific language governing permissions and
15   * limitations under the License.
16   */
17  
18  /*
19   * This is not the original file distributed by the Apache Software Foundation
20   * It has been modified by the Hipparchus project
21   */
22  package org.hipparchus.stat.correlation;
23  
24  import org.hipparchus.UnitTestUtils;
25  import org.hipparchus.exception.MathIllegalArgumentException;
26  import org.hipparchus.linear.Array2DRowRealMatrix;
27  import org.hipparchus.linear.RealMatrix;
28  import org.hipparchus.stat.descriptive.moment.Variance;
29  import org.junit.jupiter.api.Test;
30  
31  import static org.junit.jupiter.api.Assertions.assertEquals;
32  import static org.junit.jupiter.api.Assertions.fail;
33  
34  
35  class CovarianceTest {
36  
37      protected final double[] longleyData = new double[] {
38              60323,83.0,234289,2356,1590,107608,1947,
39              61122,88.5,259426,2325,1456,108632,1948,
40              60171,88.2,258054,3682,1616,109773,1949,
41              61187,89.5,284599,3351,1650,110929,1950,
42              63221,96.2,328975,2099,3099,112075,1951,
43              63639,98.1,346999,1932,3594,113270,1952,
44              64989,99.0,365385,1870,3547,115094,1953,
45              63761,100.0,363112,3578,3350,116219,1954,
46              66019,101.2,397469,2904,3048,117388,1955,
47              67857,104.6,419180,2822,2857,118734,1956,
48              68169,108.4,442769,2936,2798,120445,1957,
49              66513,110.8,444546,4681,2637,121950,1958,
50              68655,112.6,482704,3813,2552,123366,1959,
51              69564,114.2,502601,3931,2514,125368,1960,
52              69331,115.7,518173,4806,2572,127852,1961,
53              70551,116.9,554894,4007,2827,130081,1962
54          };
55  
56      protected final double[] swissData = new double[] {
57              80.2,17.0,15,12,9.96,
58              83.1,45.1,6,9,84.84,
59              92.5,39.7,5,5,93.40,
60              85.8,36.5,12,7,33.77,
61              76.9,43.5,17,15,5.16,
62              76.1,35.3,9,7,90.57,
63              83.8,70.2,16,7,92.85,
64              92.4,67.8,14,8,97.16,
65              82.4,53.3,12,7,97.67,
66              82.9,45.2,16,13,91.38,
67              87.1,64.5,14,6,98.61,
68              64.1,62.0,21,12,8.52,
69              66.9,67.5,14,7,2.27,
70              68.9,60.7,19,12,4.43,
71              61.7,69.3,22,5,2.82,
72              68.3,72.6,18,2,24.20,
73              71.7,34.0,17,8,3.30,
74              55.7,19.4,26,28,12.11,
75              54.3,15.2,31,20,2.15,
76              65.1,73.0,19,9,2.84,
77              65.5,59.8,22,10,5.23,
78              65.0,55.1,14,3,4.52,
79              56.6,50.9,22,12,15.14,
80              57.4,54.1,20,6,4.20,
81              72.5,71.2,12,1,2.40,
82              74.2,58.1,14,8,5.23,
83              72.0,63.5,6,3,2.56,
84              60.5,60.8,16,10,7.72,
85              58.3,26.8,25,19,18.46,
86              65.4,49.5,15,8,6.10,
87              75.5,85.9,3,2,99.71,
88              69.3,84.9,7,6,99.68,
89              77.3,89.7,5,2,100.00,
90              70.5,78.2,12,6,98.96,
91              79.4,64.9,7,3,98.22,
92              65.0,75.9,9,9,99.06,
93              92.2,84.6,3,3,99.46,
94              79.3,63.1,13,13,96.83,
95              70.4,38.4,26,12,5.62,
96              65.7,7.7,29,11,13.79,
97              72.7,16.7,22,13,11.22,
98              64.4,17.6,35,32,16.92,
99              77.6,37.6,15,7,4.97,
100             67.6,18.7,25,7,8.65,
101             35.0,1.2,37,53,42.34,
102             44.7,46.6,16,29,50.43,
103             42.8,27.7,22,29,58.33
104         };
105 
106 
107     /**
108      * Test Longley dataset against R.
109      * Data Source: J. Longley (1967) "An Appraisal of Least Squares
110      * Programs for the Electronic Computer from the Point of View of the User"
111      * Journal of the American Statistical Association, vol. 62. September,
112      * pp. 819-841.
113      *
114      * Data are from NIST:
115      * <a href="https://www.itl.nist.gov/div898/strd/lls/data/LINKS/DATA/Longley.dat">Longley dataset</a>
116      */
117     @Test
118     void testLongly() {
119         RealMatrix matrix = createRealMatrix(longleyData, 16, 7);
120         RealMatrix covarianceMatrix = new Covariance(matrix).getCovarianceMatrix();
121         double[] rData = new double[] {
122          12333921.73333333246, 3.679666000000000e+04, 343330206.333333313,
123          1649102.666666666744, 1117681.066666666651, 23461965.733333334, 16240.93333333333248,
124          36796.66000000000, 1.164576250000000e+02, 1063604.115416667,
125          6258.666250000000, 3490.253750000000, 73503.000000000, 50.92333333333334,
126          343330206.33333331347, 1.063604115416667e+06, 9879353659.329166412,
127          56124369.854166664183, 30880428.345833335072, 685240944.600000024, 470977.90000000002328,
128          1649102.66666666674, 6.258666250000000e+03, 56124369.854166664,
129          873223.429166666698, -115378.762499999997, 4462741.533333333, 2973.03333333333330,
130          1117681.06666666665, 3.490253750000000e+03, 30880428.345833335,
131          -115378.762499999997, 484304.095833333326, 1764098.133333333, 1382.43333333333339,
132          23461965.73333333433, 7.350300000000000e+04, 685240944.600000024,
133          4462741.533333333209, 1764098.133333333302, 48387348.933333330, 32917.40000000000146,
134          16240.93333333333, 5.092333333333334e+01, 470977.900000000,
135          2973.033333333333, 1382.433333333333, 32917.40000000, 22.66666666666667
136         };
137 
138         UnitTestUtils.customAssertEquals("covariance matrix", createRealMatrix(rData, 7, 7), covarianceMatrix, 10E-9);
139 
140     }
141 
142     /**
143      * Test R Swiss fertility dataset against R.
144      * Data Source: R datasets package
145      */
146     @Test
147     void testSwissFertility() {
148          RealMatrix matrix = createRealMatrix(swissData, 47, 5);
149          RealMatrix covarianceMatrix = new Covariance(matrix).getCovarianceMatrix();
150          double[] rData = new double[] {
151            156.0424976873265, 100.1691489361702, -64.36692876965772, -79.7295097132285, 241.5632030527289,
152            100.169148936170251, 515.7994172062905, -124.39283071230344, -139.6574005550416, 379.9043755781684,
153            -64.3669287696577, -124.3928307123034, 63.64662349676226, 53.5758556891767, -190.5606105457909,
154            -79.7295097132285, -139.6574005550416, 53.57585568917669, 92.4560592044403, -61.6988297872340,
155             241.5632030527289, 379.9043755781684, -190.56061054579092, -61.6988297872340, 1739.2945371877890
156          };
157 
158          UnitTestUtils.customAssertEquals("covariance matrix", createRealMatrix(rData, 5, 5), covarianceMatrix, 10E-13);
159     }
160 
161     /**
162      * Constant column
163      */
164     @Test
165     void testConstant() {
166         double[] noVariance = new double[] {1, 1, 1, 1};
167         double[] values = new double[] {1, 2, 3, 4};
168         assertEquals(0d, new Covariance().covariance(noVariance, values, true), Double.MIN_VALUE);
169         assertEquals(0d, new Covariance().covariance(noVariance, noVariance, true), Double.MIN_VALUE);
170     }
171 
172     /**
173      * One column
174      */
175     @Test
176     void testOneColumn() {
177         RealMatrix cov = new Covariance(new double[][] {{1}, {2}}, false).getCovarianceMatrix();
178         assertEquals(1, cov.getRowDimension());
179         assertEquals(1, cov.getColumnDimension());
180         assertEquals(0.25, cov.getEntry(0, 0), 1.0e-15);
181     }
182 
183     /**
184      * Insufficient data
185      */
186     @Test
187     void testInsufficientData() {
188         double[] one = new double[] {1};
189         double[] two = new double[] {2};
190         try {
191             new Covariance().covariance(one, two, false);
192             fail("Expecting MathIllegalArgumentException");
193         } catch (MathIllegalArgumentException ex) {
194             // Expected
195         }
196         try {
197             new Covariance(new double[][] {{},{}});
198             fail("Expecting MathIllegalArgumentException");
199         } catch (MathIllegalArgumentException ex) {
200             // Expected
201         }
202     }
203 
204     /**
205      * Verify that diagonal entries are consistent with Variance computation and matrix matches
206      * column-by-column covariances
207      */
208     @Test
209     void testConsistency() {
210         final RealMatrix matrix = createRealMatrix(swissData, 47, 5);
211         final RealMatrix covarianceMatrix = new Covariance(matrix).getCovarianceMatrix();
212 
213         // Variances on the diagonal
214         Variance variance = new Variance();
215         for (int i = 0; i < 5; i++) {
216             assertEquals(variance.evaluate(matrix.getColumn(i)), covarianceMatrix.getEntry(i,i), 10E-14);
217         }
218 
219         // Symmetry, column-consistency
220         assertEquals(covarianceMatrix.getEntry(2, 3),
221                 new Covariance().covariance(matrix.getColumn(2), matrix.getColumn(3), true), 10E-14);
222         assertEquals(covarianceMatrix.getEntry(2, 3), covarianceMatrix.getEntry(3, 2), Double.MIN_VALUE);
223 
224         // All columns same -> all entries = column variance
225         RealMatrix repeatedColumns = new Array2DRowRealMatrix(47, 3);
226         for (int i = 0; i < 3; i++) {
227             repeatedColumns.setColumnMatrix(i, matrix.getColumnMatrix(0));
228         }
229         RealMatrix repeatedCovarianceMatrix = new Covariance(repeatedColumns).getCovarianceMatrix();
230         double columnVariance = variance.evaluate(matrix.getColumn(0));
231         for (int i = 0; i < 3; i++) {
232             for (int j = 0; j < 3; j++) {
233                 assertEquals(columnVariance, repeatedCovarianceMatrix.getEntry(i, j), 10E-14);
234             }
235         }
236 
237         // Check bias-correction defaults
238         double[][] data = matrix.getData();
239         UnitTestUtils.customAssertEquals("Covariances",
240                                          covarianceMatrix, new Covariance().computeCovarianceMatrix(data), Double.MIN_VALUE);
241         UnitTestUtils.customAssertEquals("Covariances",
242                                          covarianceMatrix, new Covariance().computeCovarianceMatrix(data, true), Double.MIN_VALUE);
243 
244         double[] x = data[0];
245         double[] y = data[1];
246         assertEquals(new Covariance().covariance(x, y),
247                 new Covariance().covariance(x, y, true), Double.MIN_VALUE);
248     }
249 
250     protected RealMatrix createRealMatrix(double[] data, int nRows, int nCols) {
251         double[][] matrixData = new double[nRows][nCols];
252         int ptr = 0;
253         for (int i = 0; i < nRows; i++) {
254             System.arraycopy(data, ptr, matrixData[i], 0, nCols);
255             ptr += nCols;
256         }
257         return new Array2DRowRealMatrix(matrixData);
258     }
259 }