source: branches/StatisticalTesting/HeuristicLab.Analysis.Statistics/3.3/SampleSizeDetermination.cs @ 9706

Last change on this file since 9706 was 9706, checked in by ascheibe, 6 years ago

#2031

  • added exponential fitting
  • added logarithmic fitting
  • refactored fitting code
  • updated license headers
File size: 3.2 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2013 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
4 *
5 * This file is part of HeuristicLab.
6 *
7 * HeuristicLab is free software: you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation, either version 3 of the License, or
10 * (at your option) any later version.
11 *
12 * HeuristicLab is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
15 * GNU General Public License for more details.
16 *
17 * You should have received a copy of the GNU General Public License
18 * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
19 */
20#endregion
21
22using System;
23using System.Linq;
24using HeuristicLab.Common;
25
26namespace HeuristicLab.Analysis.Statistics {
27  public class SampleSizeDetermination {
28    /// <summary>
29    /// Determines for a given sample the sample size by estimating the means.
30    /// </summary>
31    /// <param name="samples">The pilot sample.</param>
32    /// <param name="e">Precision. </param>
33    /// <param name="conf">Confidence Interval.</param>
34    /// <returns>Number of required samples for the given confidence interval and precision. </returns>
35    public static int DetermineSampleSizeByEstimatingMean(double[] samples, double e, double conf = 0.95) {
36      if (e < 0) throw new ArgumentException("e needs to be a positive number.");
37      if (conf < 0 || conf > 1) throw new ArgumentException("The confidence Interval must be between zero and one.");
38      double result = 0;
39
40      double var = samples.StandardDeviation();
41      double n = alglib.invnormaldistribution((conf + 1) / 2);
42      result = Math.Pow(n, 2) * Math.Pow(var, 2) / Math.Pow(e, 2);
43      result = Math.Ceiling(result);
44      if (result > int.MaxValue)
45        return int.MaxValue;
46      else
47        return (int)result;
48    }
49
50    /// <summary>
51    /// Calculates Cohen's d.
52    /// </summary>
53    /// <returns>Cohen's d.
54    /// d = 0.2 means small effect
55    /// d = 0.5 means medium effect
56    /// d = 0.8 means big effect
57    /// According to Wikipedia this means: "A lower Cohen's d indicates a necessity of larger sample sizes, and vice versa."
58    /// </returns>
59    public static double CalculateCohensD(double[] d1, double[] d2) {
60      double x1, x2, s1, s2;
61
62      x1 = d1.Average();
63      x2 = d2.Average();
64      s1 = d1.Variance();
65      s2 = d2.Variance();
66
67      return (x1 - x2) / Math.Sqrt((s1 + s2) / 2);
68    }
69
70    /// <summary>
71    /// Calculates Hedges' g.
72    /// Hedges' g works like Cohen's d but corrects for bias.
73    /// </summary>
74    /// <returns>Hedges' g</returns>
75    public static double CalculateHedgesG(double[] d1, double[] d2) {
76      double x1, x2, s1, s2, n1, n2, s, g, c;
77
78      x1 = d1.Average();
79      x2 = d2.Average();
80      s1 = d1.Variance();
81      s2 = d2.Variance();
82      n1 = d1.Count();
83      n2 = d2.Count();
84
85      s = Math.Sqrt(((n1 - 1) * s1 + (n2 - 1) * s2) / (n1 + n2 - 2));
86      g = (x1 - x2) / s;
87      c = (1 - (3 / (4 * (n1 + n2) - 9))) * g;
88
89      return c;
90    }
91  }
92}
Note: See TracBrowser for help on using the repository browser.