Free cookie consent management tool by TermsFeed Policy Generator

source: branches/2817-BinPackingSpeedup/HeuristicLab.Problems.Instances.DataAnalysis/3.3/Regression/Korns/KornsFunctionNine.cs

Last change on this file was 16140, checked in by abeham, 6 years ago

#2817: updated to trunk r15680

File size: 4.5 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2018 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.Collections.Generic;
24using System.Linq;
25using HeuristicLab.Random;
26
27namespace HeuristicLab.Problems.Instances.DataAnalysis {
28  public class KornsFunctionNine : ArtificialRegressionDataDescriptor {
29
30    public override string Name { get { return "Korns 9 y = ((sqrt(X0) / log(X1)) * (exp(X2) / X3²)"; } }
31    public override string Description {
32      get {
33        return "Paper: Accuracy in Symbolic Regression" + Environment.NewLine
34        + "Authors: Michael F. Korns" + Environment.NewLine
35        + "Function: y = (sqrt(X0) / log(X1)) * (exp(X2) / X3²)" + Environment.NewLine
36        + "Binary Operators: +, -, *, % (protected division)" + Environment.NewLine
37        + "Unary Operators: sqrt, square, cube, cos, sin, tan, tanh, ln(|x|) (protected log), exp" + Environment.NewLine
38        + "Constants: random finit 64-bit IEEE double" + Environment.NewLine
39        + "\"Our testing regimen uses only statistical best practices out-of-sample testing techniques. "
40        + "We test each of the test cases on matrices of 10000 rows by 1 to 5 columns with no noise. "
41        + "For each test a training matrix is filled with random numbers between -50 and +50. The test case "
42        + "target expressions are limited to one basis function whose maximum depth is three grammar nodes.\"" + Environment.NewLine + Environment.NewLine
43        + "Note: Because of the square root and the logarithm only non-negatic values are created for their input variables!" + Environment.NewLine
44        + "Because of the exponential function only only non-positive values are created for its input variable!";
45      }
46    }
47    protected override string TargetVariable { get { return "Y"; } }
48    protected override string[] VariableNames { get { return new string[] { "X0", "X1", "X2", "X3", "X4", "Y" }; } }
49    protected override string[] AllowedInputVariables { get { return new string[] { "X0", "X1", "X2", "X3", "X4" }; } }
50    protected override int TrainingPartitionStart { get { return 0; } }
51    protected override int TrainingPartitionEnd { get { return 10000; } }
52    protected override int TestPartitionStart { get { return 10000; } }
53    protected override int TestPartitionEnd { get { return 20000; } }
54    public int Seed { get; private set; }
55
56    public KornsFunctionNine() : this((int)System.DateTime.Now.Ticks) {
57    }
58    public KornsFunctionNine(int seed) : base() {
59      Seed = seed;
60    }
61    protected override List<List<double>> GenerateValues() {
62      List<List<double>> data = new List<List<double>>();
63      var rand = new MersenneTwister((uint)Seed);
64
65      data.Add(ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 0, 50).ToList()); // note: range is only [0,50] to prevent NaN values (deviates from gp benchmark paper)
66      data.Add(ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, 0, 50).ToList()); // note: range is only [0,50] to prevent NaN values (deviates from gp benchmark paper)
67      data.Add(ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, -50, 50).ToList());
68      data.Add(ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, -50, 50).ToList());
69      data.Add(ValueGenerator.GenerateUniformDistributedValues(rand.Next(), TestPartitionEnd, -50, 50).ToList());
70
71      double x0, x1, x2, x3;
72      List<double> results = new List<double>();
73      for (int i = 0; i < data[0].Count; i++) {
74        x0 = data[0][i];
75        x1 = data[1][i];
76        x2 = data[2][i];
77        x3 = data[3][i];
78        results.Add(((Math.Sqrt(x0) / Math.Log(x1)) * (Math.Exp(x2) / Math.Pow(x3, 2))));
79      }
80      data.Add(results);
81
82      return data;
83    }
84  }
85}
Note: See TracBrowser for help on using the repository browser.