Free cookie consent management tool by TermsFeed Policy Generator

source: trunk/HeuristicLab.Problems.Instances.DataAnalysis/3.3/Regression/PennML/PennMLRegressionInstanceProvider.cs @ 15951

Last change on this file since 15951 was 15951, checked in by bburlacu, 6 years ago

#2923: Add PennML problems and implement instance provider. A reusable descriptor was also implemented, taking advantage of the structure of the data (by convention, target is named "target" and is always the last column).

File size: 3.7 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.Globalization;
25using System.IO;
26using System.IO.Compression;
27using System.Linq;
28using HeuristicLab.Data;
29
30namespace HeuristicLab.Problems.Instances.DataAnalysis.Regression.PennML {
31  public class PennMLRegressionInstanceProvider : ResourceRegressionInstanceProvider {
32    public override string Name {
33      get { return "PennML Regression Problems"; }
34    }
35
36    public override string Description {
37      get { return "A set of datasets used for benchmarking symbolic regression algorithms."; }
38    }
39
40    public override Uri WebLink {
41      get { return new Uri("https://github.com/EpistasisLab/penn-ml-benchmarks"); }
42    }
43
44    public override string ReferencePublication {
45      get { return "Patryk Orzechowski, William La Cava, Jason H. Moore - Where are we now? A large benchmark study of recent symbolic regression methods"; }
46    }
47
48    protected override string FileName {
49      get { return "PennML"; }
50    }
51
52    // the reference publication uses 75% of the samples in each of the datasets for training and the remaining 25% for testing
53    private const double trainTestSplit = 0.75;
54
55    public override IEnumerable<IDataDescriptor> GetDataDescriptors() {
56      var instanceArchiveName = GetResourceName(FileName + @"\.zip");
57      using (var instancesZipFile = new ZipArchive(GetType().Assembly.GetManifestResourceStream(instanceArchiveName), ZipArchiveMode.Read)) {
58        foreach (var entry in instancesZipFile.Entries) {
59          NumberFormatInfo numberFormat;
60          DateTimeFormatInfo dateFormat;
61          char separator;
62          using (var stream = entry.Open()) {
63            // the method below disposes the stream
64            TableFileParser.DetermineFileFormat(stream, out numberFormat, out dateFormat, out separator);
65          }
66
67          using (var stream = entry.Open()) {
68            using (var reader = new StreamReader(stream)) {
69              var header = reader.ReadLine(); // read the first line
70
71              // by convention each dataset from the PennML collection reserves the last column for the target
72              var variableNames = header.Split(separator);
73              var allowedInputVariables = variableNames.Take(variableNames.Length - 1);
74              var target = variableNames.Last();
75
76              // count lines
77              int lines = 0; while (reader.ReadLine() != null) lines++;
78
79              var trainEnd = (int)Math.Round(lines * trainTestSplit);
80              var trainRange = new IntRange(0, trainEnd);
81              var testRange = new IntRange(trainEnd, lines);
82
83              var descriptor = new PennMLRegressionDataDescriptor(entry.Name, variableNames, allowedInputVariables, target, trainRange, testRange);
84              yield return descriptor;
85            }
86          }
87        }
88      }
89    }
90  }
91}
Note: See TracBrowser for help on using the repository browser.