annotate src/main/java/edu/harvard/iq/dataverse/util/SumStatCalculator.java @ 10:a50cf11e5178

Rewrite LGDataverse completely upgrading to dataverse4.0
author Zoe Hong <zhong@mpiwg-berlin.mpg.de>
date Tue, 08 Sep 2015 17:00:21 +0200
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
10
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
1 /*
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
2 Copyright (C) 2005-2012, by the President and Fellows of Harvard College.
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
3
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
4 Licensed under the Apache License, Version 2.0 (the "License");
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
5 you may not use this file except in compliance with the License.
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
6 You may obtain a copy of the License at
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
7
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
8 http://www.apache.org/licenses/LICENSE-2.0
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
9
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
10 Unless required by applicable law or agreed to in writing, software
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
11 distributed under the License is distributed on an "AS IS" BASIS,
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
12 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
13 See the License for the specific language governing permissions and
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
14 limitations under the License.
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
15
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
16 Dataverse Network - A web application to share, preserve and analyze research data.
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
17 Developed at the Institute for Quantitative Social Science, Harvard University.
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
18 Version 3.0.
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
19 */
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
20
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
21 package edu.harvard.iq.dataverse.util;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
22 import java.util.*;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
23 import java.util.logging.Logger;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
24
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
25 import org.apache.commons.lang.*;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
26 import org.apache.commons.math.stat.*;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
27 //import cern.colt.list.*;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
28 //import cern.jet.stat.Descriptive;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
29
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
30
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
31 /**
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
32 *
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
33 * @author Leonid Andreev
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
34 */
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
35 public class SumStatCalculator {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
36
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
37 private static Logger logger = Logger.getLogger(SumStatCalculator.class.getPackage().getName());
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
38
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
39 public static double[] calculateSummaryStatistics(Number[] x){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
40 logger.fine("entering calculate summary statistics ("+x.length+" Number values);");
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
41
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
42 double[] nx = new double[8];
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
43 //("mean", "medn", "mode", "vald", "invd", "min", "max", "stdev");
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
44
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
45 Float testNanValue = new Float(Float.NaN);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
46 Number testNumberValue = testNanValue;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
47 if (Double.isNaN(testNumberValue.doubleValue())) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
48 logger.fine("Float test NaN value is still recognized as a Double NaN.");
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
49 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
50
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
51 int invalid = countInvalidValues(x);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
52 nx[4] = invalid;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
53 logger.fine("counted invalid values: "+nx[4]);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
54 nx[3] = x.length - invalid;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
55 logger.fine("counted valid values: "+nx[3]);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
56
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
57
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
58 //double[] newx = prepareForSummaryStats(x);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
59 double[] newx = prepareForSummaryStatsAlternative(x, x.length - invalid);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
60 logger.fine("prepared double vector for summary stats calculation ("+newx.length+" double values);");
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
61
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
62 ////nx[0] = StatUtils.mean(newx);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
63 nx[0] = calculateMean(newx);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
64 logger.fine("calculated mean: "+nx[0]);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
65 ////nx[1] = StatUtils.percentile(newx, 50);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
66 nx[1] = calculateMedian(newx);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
67 logger.fine("calculated medn: "+nx[1]);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
68 nx[2] = 0.0; //getMode(newx);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
69
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
70 nx[5] = StatUtils.min(newx);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
71 logger.fine("calculated min: "+nx[5]);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
72 nx[6] = StatUtils.max(newx);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
73 logger.fine("calculated max: "+nx[6]);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
74 nx[7] = Math.sqrt(StatUtils.variance(newx));
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
75 logger.fine("calculated stdev: "+nx[7]);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
76 return nx;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
77 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
78
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
79 private static double[] prepareForSummaryStats(Number[] x) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
80 Double[] z = numberToDouble(x);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
81 return removeInvalidValues(z);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
82 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
83
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
84 private static double[] prepareForSummaryStatsAlternative(Number[] x, int length) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
85 double[] retvector = new double[length];
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
86
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
87 int c = 0;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
88 for (int i = 0; i < x.length; i++) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
89 if (x[i] != null) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
90 double xvalue = x[i].doubleValue();
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
91 if (!Double.isNaN(xvalue)) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
92 retvector[c++] = xvalue;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
93 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
94 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
95 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
96
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
97 // Throw exception if c != length in the end?
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
98
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
99 return retvector;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
100 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
101
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
102 /**
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
103 * Converts an array of primitive Number types to doubles
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
104 *
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
105 */
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
106 private static Double[] numberToDouble(Number[] x){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
107 Double[] z= new Double[x.length];
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
108 for (int i=0; i<x.length;i++){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
109 z[i] = x[i] != null ? new Double( x[i].doubleValue() ) : null;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
110 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
111 return z;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
112 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
113
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
114 /**
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
115 * Returns a new double array of nulls and non-Double.NaN values only
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
116 *
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
117 */
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
118 // TODO:
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
119 // implement this in some way that does not require allocating a new
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
120 // ArrayList for the values of every vector. -- L.A. Aug. 11 2014
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
121 private static double[] removeInvalidValues(Double[] x){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
122 List<Double> dl = new ArrayList<Double>();
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
123 for (Double d : x){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
124 if (d != null && !Double.isNaN(d)){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
125 dl.add(d);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
126 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
127 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
128 return ArrayUtils.toPrimitive(
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
129 dl.toArray(new Double[dl.size()]));
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
130 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
131
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
132 /**
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
133 * Returns the number of Double.NaNs (or nulls) in a double-type array
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
134 *
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
135 */
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
136 private static int countInvalidValues(Number[] x){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
137 int counter=0;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
138 for (int i=0; i<x.length;i++){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
139 ////if ( x[i] == null || x[i].equals(Double.NaN) ) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
140 if ( x[i] == null || (Double.isNaN(x[i].doubleValue())) ) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
141 counter++;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
142 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
143 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
144 return counter;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
145 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
146
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
147 /**
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
148 * Returns the number of Double.NaNs in a double-type array
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
149 *
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
150 * TODO: figure out if this is actually necessary - to count NaNs and
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
151 * nulls separately;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
152 * -- L.A. 4.0 alpha 1
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
153 */
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
154 private static int countNaNs(double[] x){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
155 int NaNcounter=0;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
156 for (int i=0; i<x.length;i++){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
157 if (Double.isNaN(x[i])){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
158 NaNcounter++;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
159 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
160 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
161 return NaNcounter;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
162 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
163
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
164 private static double calculateMedian(double[] values) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
165 double[] sorted = new double[values.length];
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
166 System.arraycopy(values, 0, sorted, 0, values.length);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
167 logger.fine("made an extra copy of the vector;");
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
168 Arrays.sort(sorted);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
169 logger.fine("sorted double vector for median calculations;");
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
170
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
171 if (sorted.length == 0) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
172 return Double.NaN;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
173 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
174 if (sorted.length == 1) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
175 return sorted[0]; // always return single value for n = 1
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
176 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
177 double n = sorted.length;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
178 double pos = (n + 1) / 2;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
179 double fpos = Math.floor(pos);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
180 int intPos = (int) fpos;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
181 double dif = pos - fpos;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
182
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
183 double lower = sorted[intPos - 1];
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
184 double upper = sorted[intPos];
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
185
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
186 return lower + dif * (upper - lower);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
187 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
188
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
189 private static double calculateMean(double[] values) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
190 return calculateMean(values, 0 , values.length);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
191 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
192
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
193 private static double calculateMean(double[] values, final int begin, final int length) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
194
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
195 if (values == null || length == 0) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
196 return Double.NaN;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
197 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
198
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
199 double sampleSize = length;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
200
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
201 // Compute initial estimate using definitional formula
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
202 double xbar = calculateSum(values) / sampleSize;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
203
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
204 // Compute correction factor in second pass
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
205 double correction = 0;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
206 for (int i = begin; i < begin + length; i++) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
207 correction += values[i] - xbar;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
208 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
209 return xbar + (correction / sampleSize);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
210 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
211
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
212
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
213 private static double calculateSum(double[] values) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
214 return calculateSum(values, 0, values.length);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
215 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
216
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
217 private static double calculateSum(double[] values, final int begin, final int length) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
218 if (values == null || length == 0) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
219 return Double.NaN;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
220 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
221 double sum = 0.0;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
222 for (int i = begin; i < begin + length; i++) {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
223 sum += values[i];
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
224 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
225 return sum;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
226 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
227
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
228
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
229 /**
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
230 * Returns the mode statistic of a double variable
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
231 *
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
232 */
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
233 /*
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
234 public static double getMode(double[] x){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
235 double mode = Double.NaN;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
236
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
237 if ((countNaNs(x) == x.length) || (x.length < 1)){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
238 return mode;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
239 } else {
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
240 DoubleArrayList dx = new DoubleArrayList(x);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
241 dx.sort();
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
242 DoubleArrayList freqTable = new DoubleArrayList(1);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
243 IntArrayList countTable = new IntArrayList(1);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
244 Descriptive.frequencies(dx, freqTable, countTable);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
245 //out.println("freqTable="+
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
246 // ReflectionToStringBuilder.toString(freqTable));
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
247 //out.println("freqTable="+
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
248 // ReflectionToStringBuilder.toString(countTable));
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
249 int max_i = 0;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
250 for (int i=1; i< countTable.size();i++ ){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
251 if (countTable.get(i)> countTable.get(max_i)){
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
252 max_i = i;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
253 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
254 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
255 mode = freqTable.get(max_i);
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
256 //out.println("position = "+
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
257 //max_i+"\tits value="+freqTable.get(max_i));
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
258 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
259 return mode;
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
260 }
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
261 */
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
262
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
263
a50cf11e5178 Rewrite LGDataverse completely upgrading to dataverse4.0
Zoe Hong <zhong@mpiwg-berlin.mpg.de>
parents:
diff changeset
264 }