Mercurial > dive4elements > river

annotate flys-artifacts/src/main/java/de/intevation/flys/artifacts/math/GrubbsOutlier.java @ 5694:8508ee487893

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
SedimentLoadYearSelect: Add validation to have second year later than first.
author Felix Wolfsteller <felix.wolfsteller@intevation.de>
date Mon, 15 Apr 2013 10:55:08 +0200
parents a7d080347ac3
children
rev   line source
4794
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
1 package de.intevation.flys.artifacts.math;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
2
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
3 import java.util.List;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
4
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
5 import org.apache.commons.math.MathException;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
6
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
7 import org.apache.commons.math.distribution.TDistributionImpl;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
8
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
9 import org.apache.commons.math.stat.descriptive.moment.Mean;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
10 import org.apache.commons.math.stat.descriptive.moment.StandardDeviation;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
11
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
12 import org.apache.log4j.Logger;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
13
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
14 public class GrubbsOutlier
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
15 {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
16 public static final double EPSILON = 1e-5;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
17
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
18 public static final double DEFAULT_ALPHA = 0.05;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
19
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
20 private static Logger log = Logger.getLogger(GrubbsOutlier.class);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
21
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
22 protected GrubbsOutlier() {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
23 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
24
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
25 public static Integer findOutlier(List<Double> values) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
26 return findOutlier(values, DEFAULT_ALPHA, null);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
27 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
28
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
29 public static Integer findOutlier(
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
30 List<Double> values,
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
31 double alpha,
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
32 double[] stdDevResult
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
33 ) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
34 boolean debug = log.isDebugEnabled();
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
35
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
36 if (debug) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
37 log.debug("outliers significance: " + alpha);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
38 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
39
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
40 alpha = 1d - alpha;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
41
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
42 int N = values.size();
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
43
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
44 if (debug) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
45 log.debug("Values to check: " + N);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
46 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
47
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
48 if (N < 3) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
49 return null;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
50 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
51
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
52 Mean mean = new Mean();
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
53 StandardDeviation std = new StandardDeviation();
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
54
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
55 for (Double value: values) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
56 double v = value.doubleValue();
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
57 mean.increment(v);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
58 std .increment(v);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
59 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
60
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
61 double m = mean.getResult();
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
62 double s = std.getResult();
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
63
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
64 if (debug) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
65 log.debug("mean: " + m);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
66 log.debug("std dev: " + s);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
67 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
68
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
69 double maxZ = -Double.MAX_VALUE;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
70 int iv = -1;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
71 for (int i = N-1; i >= 0; --i) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
72 double v = values.get(i).doubleValue();
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
73 double z = Math.abs(v - m);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
74 if (z > maxZ) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
75 maxZ = z;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
76 iv = i;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
77 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
78 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
79
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
80 if (Math.abs(s) < EPSILON) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
81 return null;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
82 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
83
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
84 maxZ /= s;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
85
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
86 TDistributionImpl tdist = new TDistributionImpl(N-2);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
87
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
88 double t;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
89
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
90 try {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
91 t = tdist.inverseCumulativeProbability(alpha/(N+N));
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
92 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
93 catch (MathException me) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
94 log.error(me);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
95 return null;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
96 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
97
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
98 t *= t;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
99
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
100 double za = ((N-1)/Math.sqrt(N))*Math.sqrt(t/(N-2d+t));
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
101
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
102 if (debug) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
103 log.debug("max: " + maxZ + " crit: " + za);
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
104 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
105 if (stdDevResult != null) {
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
106 stdDevResult[0] = std.getResult();
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
107 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
108 return maxZ > za
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
109 ? Integer.valueOf(iv)
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
110 : null;
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
111 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
112 }
a7d080347ac3 MINFO: Allow two methods for outlier test in SQ relation.
Raimund Renkert <rrenkert@intevation.de>
parents:
diff changeset
113 // vim:set ts=4 sw=4 si et sta sts=4 fenc=utf8 :
http://dive4elements.wald.intevation.org