1 | using System;
|
---|
2 | using System.Collections;
|
---|
3 | using System.Collections.Generic;
|
---|
4 | using System.Linq;
|
---|
5 | using HeuristicLab.Common;
|
---|
6 | using HeuristicLab.Core;
|
---|
7 | using HeuristicLab.Problems.DataAnalysis;
|
---|
8 | using HeuristicLab.Random;
|
---|
9 |
|
---|
10 | namespace HeuristicLab.Problems.Instances.DataAnalysis {
|
---|
11 | public class AzzaliBenchmark1 : ArtificialRegressionDataDescriptor {
|
---|
12 | public override string Name { get { return "Azzali Benchmark1 B1 = (X4 + mean(X3)) - ((X3 · X4) - X2) ( · = dot product)"; } }
|
---|
13 | public override string Description { get { return "I. Azzali, L. Vanneschi, S. Silva, I. Bakurov, and M. Giacobini, “A Vectorial Approach to Genetic Programming,” EuroGP, pp. 213–227, 2019."; } }
|
---|
14 |
|
---|
15 | protected override string TargetVariable { get { return "B1"; } }
|
---|
16 | protected override string[] VariableNames { get { return AllowedInputVariables.Concat(new[] { TargetVariable }).ToArray(); } }
|
---|
17 | protected override string[] AllowedInputVariables { get { return new string[] { "X1", "X2", "X3", "X4" }; } }
|
---|
18 | protected override int TrainingPartitionStart { get { return 0; } }
|
---|
19 | protected override int TrainingPartitionEnd { get { return 70; } }
|
---|
20 | protected override int TestPartitionStart { get { return 70; } }
|
---|
21 | protected override int TestPartitionEnd { get { return 100; } }
|
---|
22 |
|
---|
23 | public int Seed { get; private set; }
|
---|
24 |
|
---|
25 | public AzzaliBenchmark1()
|
---|
26 | : this((int)DateTime.Now.Ticks) { }
|
---|
27 | public AzzaliBenchmark1(int seed)
|
---|
28 | : base() {
|
---|
29 | Seed = seed;
|
---|
30 | }
|
---|
31 |
|
---|
32 |
|
---|
33 | protected override List<List<double>> GenerateValues() { return null; }
|
---|
34 | protected override List<IList> GenerateValuesExtended() {
|
---|
35 | var rand = new MersenneTwister((uint)Seed);
|
---|
36 |
|
---|
37 | var x1Column = new List<double>(100);
|
---|
38 | var x2Column = new List<double>(100);
|
---|
39 | var x3Column = new List<DoubleVector>(100);
|
---|
40 | var x4Column = new List<DoubleVector>(100);
|
---|
41 | var b1Column = new List<DoubleVector>(100);
|
---|
42 |
|
---|
43 | for (int i = 0; i < 100; i++) {
|
---|
44 | var x1 = rand.NextDouble(-10, +10);
|
---|
45 | var x2 = rand.NextDouble(-10, +10);
|
---|
46 | var x3 = rand.NextDoubleVector(10, 40, 10);
|
---|
47 | var x4 = rand.NextDoubleVector(-5, +5, 10);
|
---|
48 |
|
---|
49 | var b1 = (x4 + x3.Mean()) - (x3.DotProduct(x4) - x2);
|
---|
50 |
|
---|
51 | x1Column.Add(x1);
|
---|
52 | x2Column.Add(x2);
|
---|
53 | x3Column.Add(x3);
|
---|
54 | x4Column.Add(x4);
|
---|
55 | b1Column.Add(b1);
|
---|
56 | }
|
---|
57 |
|
---|
58 | return new List<IList> { x1Column, x2Column, x3Column, x4Column, b1Column };
|
---|
59 | }
|
---|
60 | }
|
---|
61 | } |
---|