num4anova 0.0.1-java → 0.0.3-java
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +9 -5
- data/Rakefile +2 -0
- data/ext/num4anova/MultiComp.java +367 -0
- data/ext/num4anova/OneWayLayout.java +0 -1
- data/lib/dunnet.rb +89 -0
- data/lib/multicomp.rb +78 -0
- data/lib/num4anova.rb +2 -1
- metadata +5 -2
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 89f91d9c210bb998f14a726f1c7147470cba19fba721724625b7ec87e4c88ad9
|
4
|
+
data.tar.gz: a79f6e3ca4adc650193065ae62037cfed0580e5f7b9c5e1d06109d74f17a05d6
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 97a1409dfb791766292e8d6c0a49502555911529632d45b72aa6c68bc6fda7e82ed813fff4c005c9c76745a2ad05fe79d1131e23879da18afcc88d0652617c29
|
7
|
+
data.tar.gz: 4cd28879971daf44d20ff8122120327e8a74519f658a22e052724af316387c057cfb315a6d1cb23324bbd08634c066c1045594d6c8804f1f87e12d31b6acd0fb
|
data/CHANGELOG.md
CHANGED
@@ -2,13 +2,17 @@
|
|
2
2
|
|
3
3
|
## Unreleased
|
4
4
|
|
5
|
-
## [0.0.
|
5
|
+
## [0.0.3] - 2024-01-26
|
6
6
|
|
7
|
-
###
|
8
|
-
-
|
9
|
-
- chg from one-side test to two side test.
|
7
|
+
### add
|
8
|
+
- add function of dunnet_test.
|
10
9
|
|
11
|
-
## [0.0.
|
10
|
+
## [0.0.2] - 2024-01-23
|
11
|
+
|
12
|
+
### add
|
13
|
+
- add Multiple Comparisons function.
|
14
|
+
|
15
|
+
## [0.0.1] - 2024-01-10
|
12
16
|
|
13
17
|
### Fixed
|
14
18
|
- fix first fixed.
|
data/Rakefile
CHANGED
@@ -0,0 +1,367 @@
|
|
1
|
+
import org.apache.commons.math3.stat.descriptive.DescriptiveStatistics;
|
2
|
+
import java.util.Arrays;
|
3
|
+
import java.util.List;
|
4
|
+
import java.util.ArrayList;
|
5
|
+
import org.apache.commons.math3.distribution.TDistribution;
|
6
|
+
|
7
|
+
import org.apache.commons.math3.util.Combinations;
|
8
|
+
public class MultiComp {
|
9
|
+
public static class ParametrixTest {
|
10
|
+
private static ParametrixTest paramTest = new ParametrixTest();
|
11
|
+
public static ParametrixTest getInstance() {
|
12
|
+
return paramTest;
|
13
|
+
}
|
14
|
+
public boolean[][] turkeyTest(double[][] xi, double a) {
|
15
|
+
HypothesisTest hypoth = new TurkeyTest();
|
16
|
+
|
17
|
+
double[][] statistic = hypoth.calcTestStatistic(xi);
|
18
|
+
return hypoth.executeTest(statistic, a);
|
19
|
+
}
|
20
|
+
public boolean[][] bonferronoTest(double[][] xi, double a) {
|
21
|
+
HypothesisTest hypoth = new BonferroniTest();
|
22
|
+
|
23
|
+
double[][] statistic = hypoth.calcTestStatistic(xi);
|
24
|
+
return hypoth.executeTest(statistic, a * 0.5);
|
25
|
+
}
|
26
|
+
public boolean scheffe_test(double[][] xi, double a) {
|
27
|
+
return false;
|
28
|
+
}
|
29
|
+
|
30
|
+
public boolean[][] twosideTest(double[][] xi, double a) {
|
31
|
+
HypothesisTest hypoth = new TwoSideTest();
|
32
|
+
double[][] statistic = hypoth.calcTestStatistic(xi);
|
33
|
+
|
34
|
+
return hypoth.executeTest(statistic, a / 2.0);
|
35
|
+
}
|
36
|
+
public boolean[][] rightsideTest(double[][] xi, double a) {
|
37
|
+
HypothesisTest hypoth = new RightSideTest();
|
38
|
+
double[][] statistic = hypoth.calcTestStatistic(xi);
|
39
|
+
|
40
|
+
return hypoth.executeTest(statistic, a);
|
41
|
+
}
|
42
|
+
public boolean[][] leftsideTest(double[][] xi, double a) {
|
43
|
+
HypothesisTest hypoth = new LeftSideTest();
|
44
|
+
double[][] statistic = hypoth.calcTestStatistic(xi);
|
45
|
+
|
46
|
+
return hypoth.executeTest(statistic, a);
|
47
|
+
}
|
48
|
+
/*********************************/
|
49
|
+
/* interface define */
|
50
|
+
/*********************************/
|
51
|
+
private interface HypothesisTest {
|
52
|
+
double[][] calcTestStatistic(double[][] xi);
|
53
|
+
boolean[][] executeTest(double[][] statistic, double a);
|
54
|
+
}
|
55
|
+
/*********************************/
|
56
|
+
/* Class define */
|
57
|
+
/*********************************/
|
58
|
+
// turkey法
|
59
|
+
private class TurkeyTest implements HypothesisTest {
|
60
|
+
private int m = 0;
|
61
|
+
private int n = 0;
|
62
|
+
public double[][] calcTestStatistic(double[][] xi) {
|
63
|
+
n = xi.length; // a
|
64
|
+
m = xi[0].length; // n
|
65
|
+
double[][] statistic = new double[n][n];
|
66
|
+
double[] mean = calcStdMean(xi);
|
67
|
+
double ve = calcVe(xi);
|
68
|
+
|
69
|
+
for(int i = 0; i < n; i++) {
|
70
|
+
for(int j = i; j < n; j++) {
|
71
|
+
statistic[i][j] =
|
72
|
+
Math.abs(mean[i] - mean[j]) / Math.sqrt(ve / m);
|
73
|
+
}
|
74
|
+
}
|
75
|
+
|
76
|
+
return statistic;
|
77
|
+
|
78
|
+
}
|
79
|
+
public boolean[][] executeTest(double[][] statistic, double a) {
|
80
|
+
int v = n * (m - 1);
|
81
|
+
double q = qvalue(n, v, a);
|
82
|
+
boolean[][] ret = new boolean[n][n];
|
83
|
+
|
84
|
+
for(int i = 0; i < n; i++) {
|
85
|
+
for(int j = i; j < n; j++) {
|
86
|
+
ret[i][j] = (statistic[i][j] >= q) ? true : false;
|
87
|
+
}
|
88
|
+
}
|
89
|
+
return ret;
|
90
|
+
}
|
91
|
+
|
92
|
+
private double[] calcStdMean(double[][] xi) {
|
93
|
+
int n = xi.length;
|
94
|
+
double[] mean = new double[n];
|
95
|
+
|
96
|
+
for(int i = 0; i < n; i++) {
|
97
|
+
DescriptiveStatistics stat = new DescriptiveStatistics();
|
98
|
+
Arrays.stream(xi[i]).forEach(stat::addValue);
|
99
|
+
|
100
|
+
mean[i] = stat.getMean();
|
101
|
+
stat.clear();
|
102
|
+
}
|
103
|
+
return mean;
|
104
|
+
}
|
105
|
+
private double calcVe(double[][] xi) {
|
106
|
+
int n = xi.length;
|
107
|
+
int m = xi[0].length;
|
108
|
+
double na = n * (m - 1);
|
109
|
+
double sumSq = 0.0;
|
110
|
+
|
111
|
+
for(int i = 0; i < n; i++) {
|
112
|
+
DescriptiveStatistics stat = new DescriptiveStatistics();
|
113
|
+
Arrays.stream(xi[i]).forEach(stat::addValue);
|
114
|
+
double mean = stat.getMean();
|
115
|
+
for (int j = 0; j < m; j++) {
|
116
|
+
double diff = xi[i][j] - mean;
|
117
|
+
sumSq = sumSq + diff * diff;
|
118
|
+
}
|
119
|
+
stat.clear();
|
120
|
+
}
|
121
|
+
return sumSq / na;
|
122
|
+
}
|
123
|
+
// スチューデント化された範囲のa点
|
124
|
+
// v=120
|
125
|
+
// k=10 q= 4.5595 p2/a=0.016282 1/61.417516276
|
126
|
+
// k=8 q= 4.3630 p2/a=0.025263 1/39.583580731
|
127
|
+
// k=6 q= 4.0960 p2/a=0.044883 1/22.280150614
|
128
|
+
// k=5 q= 3.9169 p2/a=0.065040 1/15.375153752
|
129
|
+
// k=4 q= 3.6846 p2/a=0.103377 1/9.673331592
|
130
|
+
// k=3 q= 3.3561 p2/a=0.192270 1/5.2010194
|
131
|
+
// k=2 q= 2.8000 p2/a=0.500036 1/2
|
132
|
+
// v=14
|
133
|
+
// k=10 q= 5.2534 p2/a=0.023091 1/43.306916115
|
134
|
+
// k=8 q= 4.9903 p2/a=0.033394 1/29.945499191
|
135
|
+
// k=6 q= 4.6385 p2/a=0.054768 1/18.258837277
|
136
|
+
// k=5 q= 4.4066 p2/a=0.075883 1/13.178182202
|
137
|
+
// k=4 q= 4.1105 p2/a=0.114920 1/8.701705534
|
138
|
+
// k=3 q= 3.7014 p2/a=0.202830 1/4.930237144
|
139
|
+
// k=2 q= 3.0332 p2/a=0.499991 1/2.000036001
|
140
|
+
// v=12
|
141
|
+
// k=10 q= 5.3946 p2/a=0.024637 1/40.58935747
|
142
|
+
// k=8 q= 5.1187 p2/a=0.035180 1/28.425241615
|
143
|
+
// k=6 q= 4.7502 p2/a=0.056858 1/17.587674558
|
144
|
+
// k=5 q= 4.5077 p2/a=0.078128 1/12.799508499
|
145
|
+
// k=4 q= 4.1987 p2/a=0.117245 1/8.529148365
|
146
|
+
// k=3 q= 3.7729 p2/a=0.204923 1/4.879881712
|
147
|
+
// k=2 q= 3.0813 p2/a=0.500004 1/2
|
148
|
+
// v=10
|
149
|
+
// k=10 q= 5.5984 p2/a=0.026922 1/37.144342917
|
150
|
+
// k=8 q= 5.3042 p2/a=0.037799 1/26.455726342
|
151
|
+
// k=6 q= 4.9128 p2/a=0.059822 1/16.716258233
|
152
|
+
// k=5 q= 4.6543 p2/a=0.081351 1/12.292411894
|
153
|
+
// k=4 q= 4.3266 p2/a=0.120575 1/8.293593199
|
154
|
+
// k=3 q= 3.8768 p2/a=0.207859 1/4.810953579
|
155
|
+
// k=2 q= 3.1511 p2/a=0.499979 1/2
|
156
|
+
// v=9
|
157
|
+
// k=10 q= 5.7384 p2/a=0.028519 1/35.06434307
|
158
|
+
// k=8 q= 5.4312 p2/a=0.039639 1/25.22767981
|
159
|
+
// k=6 q= 5.0235 p2/a=0.061950 1/16.14205004
|
160
|
+
// k=5 q= 4.7554 p2/a=0.083549 1/11.969024165
|
161
|
+
// k=4 q= 4.4149 p2/a=0.122828 1/8.141466115
|
162
|
+
// k=3 q= 3.9485 p2/a=0.209848 1/4.76535397
|
163
|
+
// k=2 q= 3.1992 p2/a=0.499985 1/2
|
164
|
+
private double qvalue(int k, int v, double a) {
|
165
|
+
double den = (k-1) * (2 + k) / 2;
|
166
|
+
double p = 1.0 - a / den;
|
167
|
+
TDistribution tDist = new TDistribution(v);
|
168
|
+
double t =
|
169
|
+
tDist.inverseCumulativeProbability(p);
|
170
|
+
return Math.sqrt(2) * t;
|
171
|
+
}
|
172
|
+
}
|
173
|
+
// ボンフェロー法
|
174
|
+
private class BonferroniTest implements HypothesisTest {
|
175
|
+
private int k = 0;
|
176
|
+
private int n = 0;
|
177
|
+
private int m = 0;
|
178
|
+
public double[][] calcTestStatistic(double[][] xi) {
|
179
|
+
n = xi.length;
|
180
|
+
m = xi[0].length;
|
181
|
+
double[][] statistic = new double[n][n];
|
182
|
+
double[] mean = calcStdMean(xi);
|
183
|
+
Combinations c = new Combinations(n, 2);
|
184
|
+
List<int[]> al = new ArrayList<>();
|
185
|
+
for(int[] iterate : c) {
|
186
|
+
al.add(iterate);
|
187
|
+
}
|
188
|
+
k = al.size();
|
189
|
+
double ve = calcVe(xi);
|
190
|
+
for(int[] array : al) {
|
191
|
+
int i = array[0];
|
192
|
+
int j = array[1];
|
193
|
+
int n1 = xi[i].length;
|
194
|
+
int n2 = xi[j].length;
|
195
|
+
|
196
|
+
statistic[i][j] = Math.abs(mean[i] - mean[j])
|
197
|
+
/ Math.sqrt((1.0 / n1 + 1.0 / n2) * ve);
|
198
|
+
}
|
199
|
+
return statistic;
|
200
|
+
}
|
201
|
+
public boolean[][] executeTest(double[][] statistic, double a) {
|
202
|
+
boolean[][] ret = new boolean[n][n];
|
203
|
+
double na = n * (m - 1);
|
204
|
+
TDistribution tDist = new TDistribution(na);
|
205
|
+
|
206
|
+
double t = tDist.inverseCumulativeProbability(1.0 - a / k);
|
207
|
+
|
208
|
+
for(int i = 0; i < n; i++) {
|
209
|
+
for(int j = i; j < n; j++) {
|
210
|
+
ret[i][j] = (statistic[i][j] >= t) ? true : false;
|
211
|
+
}
|
212
|
+
}
|
213
|
+
return ret;
|
214
|
+
}
|
215
|
+
private double[] calcStdMean(double[][] xi) {
|
216
|
+
double[] mean = new double[n];
|
217
|
+
|
218
|
+
for(int i = 0; i < n; i++) {
|
219
|
+
DescriptiveStatistics stat = new DescriptiveStatistics();
|
220
|
+
Arrays.stream(xi[i]).forEach(stat::addValue);
|
221
|
+
|
222
|
+
mean[i] = stat.getMean();
|
223
|
+
stat.clear();
|
224
|
+
}
|
225
|
+
return mean;
|
226
|
+
}
|
227
|
+
private double calcVe(double[][] xi) {
|
228
|
+
double na = n * (m - 1);
|
229
|
+
double sumSq = 0.0;
|
230
|
+
|
231
|
+
for(int i = 0; i < n; i++) {
|
232
|
+
DescriptiveStatistics stat = new DescriptiveStatistics();
|
233
|
+
Arrays.stream(xi[i]).forEach(stat::addValue);
|
234
|
+
double mean = stat.getMean();
|
235
|
+
for (int j = 0; j < m; j++) {
|
236
|
+
double diff = xi[i][j] - mean;
|
237
|
+
sumSq = sumSq + diff * diff;
|
238
|
+
}
|
239
|
+
stat.clear();
|
240
|
+
}
|
241
|
+
return sumSq / na;
|
242
|
+
}
|
243
|
+
}
|
244
|
+
// ダネット法
|
245
|
+
private class DunnetTest{
|
246
|
+
private int k = 0;
|
247
|
+
private int v = 0;
|
248
|
+
private double[] mean = null;
|
249
|
+
private double[] n = null;
|
250
|
+
protected int getK() { return k;}
|
251
|
+
protected int getV() { return v;}
|
252
|
+
public double[][] calcTestStatistic(double[][] xi) {
|
253
|
+
k = xi.length;
|
254
|
+
mean = new double[k];
|
255
|
+
n = new double[k];
|
256
|
+
double[][] statistic = new double[k][k];
|
257
|
+
double ve = calcVe(xi);
|
258
|
+
|
259
|
+
for(int i = 0; i < k; i++) {
|
260
|
+
for(int j = 0; j < k; j++) {
|
261
|
+
statistic[i][j] = (mean[j] - mean[i])
|
262
|
+
/ Math.sqrt(ve * (1.0 / n[j] + 1.0 / n[i]));
|
263
|
+
}
|
264
|
+
}
|
265
|
+
return statistic;
|
266
|
+
}
|
267
|
+
private double calcVe(double[][] xi) {
|
268
|
+
double sumSq = 0.0;
|
269
|
+
int sumN = 0;
|
270
|
+
for(int i = 0; i < k; i++) {
|
271
|
+
DescriptiveStatistics stat = new DescriptiveStatistics();
|
272
|
+
Arrays.stream(xi[i]).forEach(stat::addValue);
|
273
|
+
mean[i] = stat.getMean();
|
274
|
+
n[i] = stat.getN();
|
275
|
+
sumSq += (n[i] - 1) * stat.getVariance();
|
276
|
+
sumN += n[i];
|
277
|
+
stat.clear();
|
278
|
+
}
|
279
|
+
v = sumN - k;
|
280
|
+
return sumSq / v;
|
281
|
+
}
|
282
|
+
}
|
283
|
+
private class TwoSideTest extends DunnetTest
|
284
|
+
implements HypothesisTest {
|
285
|
+
public boolean[][] executeTest(double[][] statistic, double a) {
|
286
|
+
int v = super.getV();
|
287
|
+
int k = super.getK();
|
288
|
+
double den = k - 1;
|
289
|
+
double p = 1.0 - a / den;
|
290
|
+
TDistribution tDist = new TDistribution(v);
|
291
|
+
double l_val = tDist.inverseCumulativeProbability(a / den);
|
292
|
+
double r_val = tDist.inverseCumulativeProbability(1.0 - a / den);
|
293
|
+
boolean[][] ret = new boolean[k][k];
|
294
|
+
|
295
|
+
for(int i = 0; i < k; i++) {
|
296
|
+
for(int j = 0; j < k; j++) {
|
297
|
+
ret[i][j] = evaluation(statistic[i][j], l_val, r_val );
|
298
|
+
}
|
299
|
+
}
|
300
|
+
return ret;
|
301
|
+
}
|
302
|
+
private boolean evaluation(double statistic, double l_val, double r_val) {
|
303
|
+
boolean ret = true;
|
304
|
+
|
305
|
+
if ((l_val < statistic) && (statistic < r_val)) {
|
306
|
+
ret = false;
|
307
|
+
}
|
308
|
+
return ret;
|
309
|
+
}
|
310
|
+
}
|
311
|
+
private class RightSideTest extends DunnetTest
|
312
|
+
implements HypothesisTest {
|
313
|
+
public boolean[][] executeTest(double[][] statistic, double a) {
|
314
|
+
int v = super.getV();
|
315
|
+
int k = super.getK();
|
316
|
+
double den = k - 1;
|
317
|
+
double p = 1.0 - a / den;
|
318
|
+
TDistribution tDist = new TDistribution(v);
|
319
|
+
double r_val = tDist.inverseCumulativeProbability(1.0 - a);
|
320
|
+
boolean[][] ret = new boolean[k][k];
|
321
|
+
|
322
|
+
for(int i = 0; i < k; i++) {
|
323
|
+
for(int j = 0; j < k; j++) {
|
324
|
+
ret[i][j] = evaluation(statistic[i][j], r_val );
|
325
|
+
}
|
326
|
+
}
|
327
|
+
return ret;
|
328
|
+
}
|
329
|
+
private boolean evaluation(double statistic, double r_val) {
|
330
|
+
boolean ret = true;
|
331
|
+
|
332
|
+
if (statistic < r_val) {
|
333
|
+
ret = false;
|
334
|
+
}
|
335
|
+
return ret;
|
336
|
+
}
|
337
|
+
}
|
338
|
+
private class LeftSideTest extends DunnetTest
|
339
|
+
implements HypothesisTest {
|
340
|
+
public boolean[][] executeTest(double[][] statistic, double a) {
|
341
|
+
int v = super.getV();
|
342
|
+
int k = super.getK();
|
343
|
+
double den = k - 1;
|
344
|
+
double p = a / den;
|
345
|
+
TDistribution tDist = new TDistribution(v);
|
346
|
+
double l_val = tDist.inverseCumulativeProbability(a);
|
347
|
+
boolean[][] ret = new boolean[k][k];
|
348
|
+
|
349
|
+
for(int i = 0; i < k; i++) {
|
350
|
+
for(int j = 0; j < k; j++) {
|
351
|
+
ret[i][j] = evaluation(statistic[i][j], l_val );
|
352
|
+
}
|
353
|
+
}
|
354
|
+
return ret;
|
355
|
+
}
|
356
|
+
private boolean evaluation(double statistic, double l_val) {
|
357
|
+
boolean ret = true;
|
358
|
+
|
359
|
+
if (l_val < statistic) {
|
360
|
+
ret = false;
|
361
|
+
}
|
362
|
+
return ret;
|
363
|
+
}
|
364
|
+
}
|
365
|
+
}
|
366
|
+
}
|
367
|
+
|
@@ -17,7 +17,6 @@ import org.jfree.chart.axis.CategoryAxis;
|
|
17
17
|
import org.jfree.chart.axis.NumberAxis;
|
18
18
|
import org.jfree.chart.renderer.category.LineAndShapeRenderer;
|
19
19
|
import org.jfree.chart.labels.StandardCategoryToolTipGenerator;
|
20
|
-
import org.jfree.chart.renderer.xy.XYLineAndShapeRenderer;
|
21
20
|
|
22
21
|
import org.jfree.chart.ChartUtils;
|
23
22
|
import java.io.File;
|
data/lib/dunnet.rb
ADDED
@@ -0,0 +1,89 @@
|
|
1
|
+
# Dunnet検定
|
2
|
+
# (Apache commoms math3使用)
|
3
|
+
module DunnetTestLib
|
4
|
+
# Dunnet検定の両側検定
|
5
|
+
#
|
6
|
+
# @overload twoside_test(xi, a)
|
7
|
+
# @param [array] xi データ(double[][])
|
8
|
+
# @param [double] a 有意水準
|
9
|
+
# @return [Array] 検定結果(boolean[][] true:棄却域内 false:棄却域外)
|
10
|
+
# @example
|
11
|
+
# xi = [
|
12
|
+
# [12.2, 18.8, 18.2],
|
13
|
+
# [22.2, 20.5, 14.6],
|
14
|
+
# [20.8, 19.5, 26.3],
|
15
|
+
# [26.4, 32.5, 31.3],
|
16
|
+
# [24.5, 21.2, 22.4],
|
17
|
+
# ]
|
18
|
+
# paraTest = MultiCompLib::ParametrixTestLib.new
|
19
|
+
# paraTest.twoside_test(xi, 0.05)
|
20
|
+
# =>
|
21
|
+
# res = [
|
22
|
+
# [false, false, false, true, false],
|
23
|
+
# [false, false, false, true, false],
|
24
|
+
# [false, false, false, false, false],
|
25
|
+
# [true, true, false, false, false],
|
26
|
+
# [false, false, false, false, false],
|
27
|
+
# ]
|
28
|
+
def twoside_test(xi, a)
|
29
|
+
ret = @paramTest.twosideTest(xi.to_java(Java::double[]), a)
|
30
|
+
return ret.to_a
|
31
|
+
end
|
32
|
+
# Dunnet検定の右側検定
|
33
|
+
#
|
34
|
+
# @overload rightside_test(xi, a)
|
35
|
+
# @param [array] xi データ(double[][])
|
36
|
+
# @param [double] a 有意水準
|
37
|
+
# @return [Array] 検定結果(boolean[][] true:棄却域内 false:棄却域外)
|
38
|
+
# @example
|
39
|
+
# xi = [
|
40
|
+
# [12.2, 18.8, 18.2],
|
41
|
+
# [22.2, 20.5, 14.6],
|
42
|
+
# [20.8, 19.5, 26.3],
|
43
|
+
# [26.4, 32.5, 31.3],
|
44
|
+
# [24.5, 21.2, 22.4],
|
45
|
+
# ]
|
46
|
+
# paraTest = MultiCompLib::ParametrixTestLib.new
|
47
|
+
# paraTest.rightside_test(xi, 0.05)
|
48
|
+
# =>
|
49
|
+
# res = [
|
50
|
+
# [false, false, true, true, true],
|
51
|
+
# [false, false, false, true, false],
|
52
|
+
# [false, false, false, true, false],
|
53
|
+
# [false, false, false, false, false],
|
54
|
+
# [false, false, false, true, false],
|
55
|
+
# ]
|
56
|
+
def rightside_test(xi, a)
|
57
|
+
ret = @paramTest.rightsideTest(xi.to_java(Java::double[]), a)
|
58
|
+
return ret.to_a
|
59
|
+
end
|
60
|
+
# Dunnet検定の左側検定
|
61
|
+
#
|
62
|
+
# @overload leftside_test(xi, a)
|
63
|
+
# @param [array] xi データ(double[][])
|
64
|
+
# @param [double] a 有意水準
|
65
|
+
# @return [Array] 検定結果(boolean[][] true:棄却域内 false:棄却域外)
|
66
|
+
# @example
|
67
|
+
# xi = [
|
68
|
+
# [12.2, 18.8, 18.2],
|
69
|
+
# [22.2, 20.5, 14.6],
|
70
|
+
# [20.8, 19.5, 26.3],
|
71
|
+
# [26.4, 32.5, 31.3],
|
72
|
+
# [24.5, 21.2, 22.4],
|
73
|
+
# ]
|
74
|
+
# paraTest = MultiCompLib::ParametrixTestLib.new
|
75
|
+
# paraTest.leftside_test(xi, 0.05)
|
76
|
+
# =>
|
77
|
+
# res = [
|
78
|
+
# [false, false, false, false, false],
|
79
|
+
# [false, false, false, false, false],
|
80
|
+
# [true, false, false, false, false],
|
81
|
+
# [true, true, true, false, true],
|
82
|
+
# [true, false, false, false, false],
|
83
|
+
# ]
|
84
|
+
def leftside_test(xi, a)
|
85
|
+
ret = @paramTest.leftsideTest(xi.to_java(Java::double[]), a)
|
86
|
+
return ret.to_a
|
87
|
+
end
|
88
|
+
end
|
89
|
+
|
data/lib/multicomp.rb
ADDED
@@ -0,0 +1,78 @@
|
|
1
|
+
require 'java'
|
2
|
+
require 'num4anova.jar'
|
3
|
+
require 'jfreechart-1.5.4.jar'
|
4
|
+
require 'commons-math3-3.6.1.jar'
|
5
|
+
|
6
|
+
require_relative('dunnet')
|
7
|
+
|
8
|
+
java_import 'MultiComp'
|
9
|
+
# 多重比較を行う
|
10
|
+
# (Apache commoms math3使用)
|
11
|
+
module MultiCompLib
|
12
|
+
# パラメトリック検定
|
13
|
+
class ParametrixTestLib
|
14
|
+
include DunnetTestLib
|
15
|
+
def initialize
|
16
|
+
@paramTest = MultiComp::ParametrixTest.getInstance()
|
17
|
+
end
|
18
|
+
# turkeyの方法による多重比較
|
19
|
+
#
|
20
|
+
# @overload turkey_test(xi, a)
|
21
|
+
# @param [array] xi データ(double[][])
|
22
|
+
# @param [double] a 有意水準
|
23
|
+
# @return [Array] 検定結果(boolean[][] true:棄却域内 false:棄却域外)
|
24
|
+
# @example
|
25
|
+
# xi = [
|
26
|
+
# [12.2, 18.8, 18.2],
|
27
|
+
# [22.2, 20.5, 14.6],
|
28
|
+
# [20.8, 19.5, 26.3],
|
29
|
+
# [26.4, 32.5, 31.3],
|
30
|
+
# [24.5, 21.2, 22.4],
|
31
|
+
# ]
|
32
|
+
# paraTest = MultiCompLib::ParametrixTestLib.new
|
33
|
+
# paraTest.turkey_test(xi, 0.05)
|
34
|
+
# =>
|
35
|
+
# [
|
36
|
+
# [false, false, false, true, false],
|
37
|
+
# [false, false, false, true, false],
|
38
|
+
# [false, false, false, false, false],
|
39
|
+
# [false, false, false, false, false],
|
40
|
+
# [false, false, false, false, false],
|
41
|
+
# ]
|
42
|
+
def turkey_test(xi, a)
|
43
|
+
ret = @paramTest.turkeyTest(xi.to_java(Java::double[]), a)
|
44
|
+
return ret.to_a
|
45
|
+
end
|
46
|
+
# ボンフェロー二の不等式による多重比較
|
47
|
+
#
|
48
|
+
# @overload bonferrono_test(xi, a)
|
49
|
+
# @param [array] xi データ(double[][])
|
50
|
+
# @param [double] a 有意水準
|
51
|
+
# @return [Array] 検定結果(boolean[][] true:棄却域内 false:棄却域外)
|
52
|
+
# @example
|
53
|
+
# xi = [
|
54
|
+
# [12.2, 18.8, 18.2],
|
55
|
+
# [22.2, 20.5, 14.6],
|
56
|
+
# [20.8, 19.5, 26.3],
|
57
|
+
# [26.4, 32.5, 31.3],
|
58
|
+
# [24.5, 21.2, 22.4],
|
59
|
+
# ]
|
60
|
+
# paraTest = MultiCompLib::ParametrixTestLib.new
|
61
|
+
# paraTest.bonferrono_test(xi, 0.05)
|
62
|
+
# =>
|
63
|
+
# [
|
64
|
+
# [false, false, false, true, false],
|
65
|
+
# [false, false, false, true, false],
|
66
|
+
# [false, false, false, false, false],
|
67
|
+
# [false, false, false, false, false],
|
68
|
+
# [false, false, false, false, false],
|
69
|
+
# ]
|
70
|
+
def bonferrono_test(xi, a)
|
71
|
+
ret = @paramTest.bonferronoTest(xi.to_java(Java::double[]), a)
|
72
|
+
return ret.to_a
|
73
|
+
end
|
74
|
+
end
|
75
|
+
# ノンパラメトリック検定
|
76
|
+
class NonParametrixTestLib
|
77
|
+
end
|
78
|
+
end
|
data/lib/num4anova.rb
CHANGED
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: num4anova
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.0.
|
4
|
+
version: 0.0.3
|
5
5
|
platform: java
|
6
6
|
authors:
|
7
7
|
- siranovel
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2024-01-
|
11
|
+
date: 2024-01-26 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: rake
|
@@ -61,10 +61,13 @@ files:
|
|
61
61
|
- Gemfile
|
62
62
|
- LICENSE
|
63
63
|
- Rakefile
|
64
|
+
- ext/num4anova/MultiComp.java
|
64
65
|
- ext/num4anova/OneWayLayout.java
|
65
66
|
- lib/commons-math3-3.6.1.jar
|
67
|
+
- lib/dunnet.rb
|
66
68
|
- lib/jcommon-1.0.23.jar
|
67
69
|
- lib/jfreechart-1.5.4.jar
|
70
|
+
- lib/multicomp.rb
|
68
71
|
- lib/num4anova.rb
|
69
72
|
homepage: http://github.com/siranovel/num4varanly
|
70
73
|
licenses:
|