1 | #region License Information
|
---|
2 | /* HeuristicLab
|
---|
3 | * Copyright (C) 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 |
|
---|
22 | using System;
|
---|
23 | using System.Collections;
|
---|
24 | using System.Collections.Generic;
|
---|
25 | using System.Linq;
|
---|
26 | using HeuristicLab.Core;
|
---|
27 | using HeuristicLab.Random;
|
---|
28 | using MathNet.Numerics.Statistics;
|
---|
29 |
|
---|
30 | using DoubleVector = MathNet.Numerics.LinearAlgebra.Vector<double>;
|
---|
31 |
|
---|
32 | namespace HeuristicLab.Problems.Instances.DataAnalysis {
|
---|
33 | public abstract class VectorDataTestOne : ArtificialRegressionDataDescriptor {
|
---|
34 |
|
---|
35 | protected const int Rows = 100;
|
---|
36 |
|
---|
37 | public override string Description { get { return ""; } }
|
---|
38 |
|
---|
39 | protected override string TargetVariable { get { return "Y"; } }
|
---|
40 | protected override string[] VariableNames { get { return new string[] { "X1", "X2", "X3", "V1", "V2", "Y" }; } }
|
---|
41 | protected override string[] AllowedInputVariables { get { return new string[] { "X1", "X2", "X3", "V1", "V2" }; } }
|
---|
42 | protected override int TrainingPartitionStart { get { return 0; } }
|
---|
43 | protected override int TrainingPartitionEnd { get { return Rows * 3 / 4; } }
|
---|
44 | protected override int TestPartitionStart { get { return TrainingPartitionEnd; } }
|
---|
45 | protected override int TestPartitionEnd { get { return Rows; } }
|
---|
46 |
|
---|
47 | public int Seed { get; private set; }
|
---|
48 |
|
---|
49 | protected VectorDataTestOne()
|
---|
50 | : this((int)DateTime.Now.Ticks) { }
|
---|
51 | protected VectorDataTestOne(int seed)
|
---|
52 | : base() {
|
---|
53 | Seed = seed;
|
---|
54 | }
|
---|
55 |
|
---|
56 |
|
---|
57 | protected override List<List<double>> GenerateValues() { return null; }
|
---|
58 | protected override List<IList> GenerateValuesExtended() {
|
---|
59 | var rand = new MersenneTwister((uint)Seed);
|
---|
60 |
|
---|
61 | double x1, x2, x3;
|
---|
62 | DoubleVector v1, v2;
|
---|
63 | double y;
|
---|
64 |
|
---|
65 | var x1Column = new List<double>(Rows);
|
---|
66 | var x2Column = new List<double>(Rows);
|
---|
67 | var x3Column = new List<double>(Rows);
|
---|
68 | var v1Column = new List<DoubleVector>(Rows);
|
---|
69 | var v2Column = new List<DoubleVector>(Rows);
|
---|
70 | var yColumn = new List<double>(Rows);
|
---|
71 |
|
---|
72 | var vectorLengths = GetVectorLengths(rand);
|
---|
73 | for (int i = 0; i < Rows; i++) {
|
---|
74 | x1 = rand.NextDouble(-2, 2);
|
---|
75 | x2 = rand.NextDouble(2, 6);
|
---|
76 | x3 = rand.NextDouble(0, 1);
|
---|
77 | int v1Length = vectorLengths[0][i];
|
---|
78 | int v2Length = vectorLengths[1][i];
|
---|
79 | v1 = rand.NextDoubleVector(-2, 2, v1Length);
|
---|
80 | v2 = rand.NextDoubleVector(3, 5, v2Length);
|
---|
81 |
|
---|
82 | y = x1 * v1.Sum() + x2 * v2.Mean();
|
---|
83 |
|
---|
84 | x1Column.Add(x1);
|
---|
85 | x2Column.Add(x2);
|
---|
86 | x3Column.Add(x3);
|
---|
87 | v1Column.Add(v1);
|
---|
88 | v2Column.Add(v2);
|
---|
89 | yColumn.Add(y);
|
---|
90 | }
|
---|
91 |
|
---|
92 | return new List<IList> { x1Column, x2Column, x3Column, v1Column, v2Column, yColumn };
|
---|
93 | }
|
---|
94 |
|
---|
95 | protected abstract List<int>[] GetVectorLengths(IRandom rand);
|
---|
96 | }
|
---|
97 |
|
---|
98 | public class VectorDataTestOneA : VectorDataTestOne {
|
---|
99 | public override string Name {
|
---|
100 | get { return "Vector Data Test - I [fully-constrained]: Y = X1 * sum(V1) + X2 * mean(V2)"; }
|
---|
101 | }
|
---|
102 |
|
---|
103 | public VectorDataTestOneA() : base() { }
|
---|
104 | public VectorDataTestOneA(int seed) : base(seed) { }
|
---|
105 |
|
---|
106 | protected override List<int>[] GetVectorLengths(IRandom rand) {
|
---|
107 | // always same length
|
---|
108 | const int length = 5;
|
---|
109 | return new List<int>[2] {
|
---|
110 | Enumerable.Repeat(length, Rows).ToList(),
|
---|
111 | Enumerable.Repeat(length, Rows).ToList()
|
---|
112 | };
|
---|
113 | }
|
---|
114 | }
|
---|
115 |
|
---|
116 | public class VectorDataTestOneB : VectorDataTestOne {
|
---|
117 | public override string Name { get { return "Vector Data Test - I [row-constrained]: Y = X1 * sum(V1) + X2 * mean(V2)"; } }
|
---|
118 |
|
---|
119 | public VectorDataTestOneB() : base() { }
|
---|
120 | public VectorDataTestOneB(int seed) : base(seed) { }
|
---|
121 |
|
---|
122 | protected override List<int>[] GetVectorLengths(IRandom rand) {
|
---|
123 | // length between length 4 and 8, same row always the same length
|
---|
124 | var lengths = Enumerable.Range(0, Rows).Select(i => rand.Next(4, 8)).ToList();
|
---|
125 | return new List<int>[2] {
|
---|
126 | Enumerable.Range(0, Rows).Select(i => lengths[i]).ToList(),
|
---|
127 | Enumerable.Range(0, Rows).Select(i => lengths[i]).ToList()
|
---|
128 | };
|
---|
129 | }
|
---|
130 | }
|
---|
131 |
|
---|
132 | public class VectorDataTestOneC : VectorDataTestOne {
|
---|
133 | public override string Name { get { return "Vector Data Test - I [column-constrained]: Y = X1 * sum(V1) + X2 * mean(V2)"; } }
|
---|
134 |
|
---|
135 | public VectorDataTestOneC() : base() { }
|
---|
136 | public VectorDataTestOneC(int seed) : base(seed) { }
|
---|
137 |
|
---|
138 | protected override List<int>[] GetVectorLengths(IRandom rand) {
|
---|
139 | // length between length 4 and 8; each feature is same length
|
---|
140 | // force two different lengths
|
---|
141 | int v1Length = rand.Next(4, 8);
|
---|
142 | int v2Length;
|
---|
143 | do {
|
---|
144 | v2Length = rand.Next(4, 8);
|
---|
145 | } while (v1Length != v2Length);
|
---|
146 | return new List<int>[2] {
|
---|
147 | Enumerable.Repeat(v1Length, Rows).ToList(),
|
---|
148 | Enumerable.Repeat(v2Length, Rows).ToList()
|
---|
149 | };
|
---|
150 | }
|
---|
151 | }
|
---|
152 |
|
---|
153 | public class VectorDataTestOneD : VectorDataTestOne {
|
---|
154 | public override string Name { get { return "Vector Data Test - I [unconstrained]: Y = X1 * sum(V1) + X2 * mean(V2)"; } }
|
---|
155 |
|
---|
156 | public VectorDataTestOneD() : base() { }
|
---|
157 | public VectorDataTestOneD(int seed) : base(seed) { }
|
---|
158 |
|
---|
159 | protected override List<int>[] GetVectorLengths(IRandom rand) {
|
---|
160 | // always random between 4 and 8
|
---|
161 | return new List<int>[2] {
|
---|
162 | Enumerable.Range(0, Rows).Select(i => rand.Next(4, 8)).ToList(),
|
---|
163 | Enumerable.Range(0, Rows).Select(i => rand.Next(4, 8)).ToList()
|
---|
164 | };
|
---|
165 | }
|
---|
166 | }
|
---|
167 | } |
---|