source: trunk/sources/HeuristicLab.Problems.DataAnalysis/3.3/DataAnalysisProblemData.cs @ 5013

Last change on this file since 5013 was 5013, checked in by gkronber, 12 years ago

Implemented heuristic to determine format for import of data tables and test cases. #1173

File size: 20.8 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2010 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.IO;
25using System.Linq;
26using HeuristicLab.Common;
27using HeuristicLab.Core;
28using HeuristicLab.Data;
29using HeuristicLab.Parameters;
30using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
31
32namespace HeuristicLab.Problems.DataAnalysis {
33  [Item("DataAnalysisProblemData", "Represents an item containing all data defining a data analysis problem.")]
34  [StorableClass]
35  public class DataAnalysisProblemData : ParameterizedNamedItem, IStorableContent {
36    protected bool suppressEvents = false;
37    #region IStorableContent Members
38    public string Filename { get; set; }
39    #endregion
40    #region default data
41    // y = x^4 + x^3 + x^2 + x
42    private static double[,] kozaF1 = new double[,] {
43{2.017885919, -1.449165046},
44{1.30060506,  -1.344523885},
45{1.147134798, -1.317989331},
46{0.877182504, -1.266142284},
47{0.852562452, -1.261020794},
48{0.431095788, -1.158793317},
49{0.112586002, -1.050908405},
50{0.04594507,  -1.021989402},
51{0.042572879, -1.020438113},
52{-0.074027291,  -0.959859562},
53{-0.109178553,  -0.938094706},
54{-0.259721109,  -0.803635355},
55{-0.272991057,  -0.387519561},
56{-0.161978191,  -0.193611001},
57{-0.102489983,  -0.114215349},
58{-0.01469968, -0.014918985},
59{-0.008863365,  -0.008942626},
60{0.026751057, 0.026054094},
61{0.166922436, 0.14309643},
62{0.176953808, 0.1504144},
63{0.190233418, 0.159916534},
64{0.199800708, 0.166635331},
65{0.261502822, 0.207600348},
66{0.30182879,  0.232370249},
67{0.83763905,  0.468046718}
68    };
69    #endregion
70    #region parameter properties
71    public ValueParameter<Dataset> DatasetParameter {
72      get { return (ValueParameter<Dataset>)Parameters["Dataset"]; }
73    }
74    public IValueParameter<StringValue> TargetVariableParameter {
75      get { return (IValueParameter<StringValue>)Parameters["TargetVariable"]; }
76    }
77    public IValueParameter<ICheckedItemList<StringValue>> InputVariablesParameter {
78      get { return (IValueParameter<ICheckedItemList<StringValue>>)Parameters["InputVariables"]; }
79    }
80    public IValueParameter<IntValue> TrainingSamplesStartParameter {
81      get { return (IValueParameter<IntValue>)Parameters["TrainingSamplesStart"]; }
82    }
83    public IValueParameter<IntValue> TrainingSamplesEndParameter {
84      get { return (IValueParameter<IntValue>)Parameters["TrainingSamplesEnd"]; }
85    }
86    public IValueParameter<IntValue> TestSamplesStartParameter {
87      get { return (IValueParameter<IntValue>)Parameters["TestSamplesStart"]; }
88    }
89    public IValueParameter<IntValue> TestSamplesEndParameter {
90      get { return (IValueParameter<IntValue>)Parameters["TestSamplesEnd"]; }
91    }
92    public IValueParameter<PercentValue> ValidationPercentageParameter {
93      get { return (IValueParameter<PercentValue>)Parameters["ValidationPercentage"]; }
94    }
95    #endregion
96
97    #region properties
98    public Dataset Dataset {
99      get { return DatasetParameter.Value; }
100      set {
101        if (value != Dataset) {
102          if (value == null) throw new ArgumentNullException();
103          DatasetParameter.Value = value;
104        }
105      }
106    }
107    public StringValue TargetVariable {
108      get { return TargetVariableParameter.Value; }
109      set {
110        if (value != TargetVariableParameter.Value) {
111          if (value == null) throw new ArgumentNullException();
112          if (TargetVariable != null) DeregisterStringValueEventHandlers(TargetVariable);
113          TargetVariableParameter.Value = value;
114        }
115      }
116    }
117    public ICheckedItemList<StringValue> InputVariables {
118      get { return InputVariablesParameter.Value; }
119      set {
120        if (value != InputVariables) {
121          if (value == null) throw new ArgumentNullException();
122          if (InputVariables != null) DeregisterInputVariablesEventHandlers();
123          InputVariablesParameter.Value = value;
124        }
125      }
126    }
127    public IntValue TrainingSamplesStart {
128      get { return TrainingSamplesStartParameter.Value; }
129      set {
130        if (value != TrainingSamplesStart) {
131          if (value == null) throw new ArgumentNullException();
132          if (TrainingSamplesStart != null) DeregisterValueTypeEventHandlers(TrainingSamplesStart);
133          TrainingSamplesStartParameter.Value = value;
134        }
135      }
136    }
137    public IntValue TrainingSamplesEnd {
138      get { return TrainingSamplesEndParameter.Value; }
139      set {
140        if (value != TrainingSamplesEnd) {
141          if (value == null) throw new ArgumentNullException();
142          if (TrainingSamplesEnd != null) DeregisterValueTypeEventHandlers(TrainingSamplesEnd);
143          TrainingSamplesEndParameter.Value = value;
144        }
145      }
146    }
147    public IntValue TestSamplesStart {
148      get { return TestSamplesStartParameter.Value; }
149      set {
150        if (value != TestSamplesStart) {
151          if (value == null) throw new ArgumentNullException();
152          if (TestSamplesStart != null) DeregisterValueTypeEventHandlers(TestSamplesStart);
153          TestSamplesStartParameter.Value = value;
154        }
155      }
156    }
157    public IntValue TestSamplesEnd {
158      get { return TestSamplesEndParameter.Value; }
159      set {
160        if (value != TestSamplesEnd) {
161          if (value == null) throw new ArgumentNullException();
162          if (TestSamplesEnd != null) DeregisterValueTypeEventHandlers(TestSamplesEnd);
163          TestSamplesEndParameter.Value = value;
164        }
165      }
166    }
167    public PercentValue ValidationPercentage {
168      get { return ValidationPercentageParameter.Value; }
169      set {
170        if (value != ValidationPercentage) {
171          if (value == null) throw new ArgumentNullException();
172          if (value.Value < 0 || value.Value > 1) throw new ArgumentException("ValidationPercentage must be between 0 and 1.");
173          if (ValidationPercentage != null) DeregisterValueTypeEventHandlers(ValidationPercentage);
174          ValidationPercentageParameter.Value = value;
175        }
176      }
177    }
178
179    public IEnumerable<int> TrainingIndizes {
180      get {
181        return Enumerable.Range(TrainingSamplesStart.Value, TrainingSamplesEnd.Value - TrainingSamplesStart.Value)
182                         .Where(i => i >= 0 && i < Dataset.Rows && (i < TestSamplesStart.Value || TestSamplesEnd.Value <= i));
183      }
184    }
185    public IEnumerable<int> TestIndizes {
186      get {
187        return Enumerable.Range(TestSamplesStart.Value, TestSamplesEnd.Value - TestSamplesStart.Value)
188           .Where(i => i >= 0 && i < Dataset.Rows);
189      }
190    }
191    #endregion
192
193    [StorableConstructor]
194    protected DataAnalysisProblemData(bool deserializing) : base(deserializing) { }
195    protected DataAnalysisProblemData(DataAnalysisProblemData original, Cloner cloner)
196      : base(original, cloner) {
197      RegisterParameterEventHandlers();
198      RegisterParameterValueEventHandlers();
199    }
200    public DataAnalysisProblemData()
201      : base() {
202      var inputVariables = new CheckedItemList<StringValue>();
203      StringValue inputVariable = new StringValue("x");
204      inputVariables.Add(inputVariable);
205      StringValue targetVariable = new StringValue("y");
206      var validTargetVariables = new ItemSet<StringValue>();
207      validTargetVariables.Add(targetVariable);
208      Parameters.Add(new ValueParameter<Dataset>("Dataset", new Dataset(new string[] { "y", "x" }, kozaF1)));
209      Parameters.Add(new ValueParameter<ICheckedItemList<StringValue>>("InputVariables", inputVariables.AsReadOnly()));
210      Parameters.Add(new ConstrainedValueParameter<StringValue>("TargetVariable", validTargetVariables, targetVariable));
211      Parameters.Add(new ValueParameter<IntValue>("TrainingSamplesStart", new IntValue(0)));
212      Parameters.Add(new ValueParameter<IntValue>("TrainingSamplesEnd", new IntValue(15)));
213      Parameters.Add(new ValueParameter<IntValue>("TestSamplesStart", new IntValue(15)));
214      Parameters.Add(new ValueParameter<IntValue>("TestSamplesEnd", new IntValue(25)));
215      Parameters.Add(new ValueParameter<PercentValue>("ValidationPercentage", "The relative amount of the training samples that should be used as validation set.", new PercentValue(0.5)));
216
217      DatasetParameter.ReactOnValueToStringChangedAndValueItemImageChanged = false;
218      RegisterParameterEventHandlers();
219      RegisterParameterValueEventHandlers();
220    }
221
222    public DataAnalysisProblemData(Dataset dataset, IEnumerable<string> inputVariables, string targetVariable,
223      int trainingSamplesStart, int trainingSamplesEnd, int testSamplesStart, int testSamplesEnd) {
224      var inputVariablesList = new CheckedItemList<StringValue>(inputVariables.Select(x => new StringValue(x)).ToList());
225      StringValue targetVariableValue = new StringValue(targetVariable);
226      var validTargetVariables = new ItemSet<StringValue>();
227      foreach (var variable in dataset.VariableNames)
228        if (variable != targetVariable)
229          validTargetVariables.Add(new StringValue(variable));
230      validTargetVariables.Add(targetVariableValue);
231      Parameters.Add(new ValueParameter<Dataset>("Dataset", dataset));
232      Parameters.Add(new ValueParameter<ICheckedItemList<StringValue>>("InputVariables", inputVariablesList.AsReadOnly()));
233      Parameters.Add(new ConstrainedValueParameter<StringValue>("TargetVariable", validTargetVariables, targetVariableValue));
234      Parameters.Add(new ValueParameter<IntValue>("TrainingSamplesStart", new IntValue(trainingSamplesStart)));
235      Parameters.Add(new ValueParameter<IntValue>("TrainingSamplesEnd", new IntValue(trainingSamplesEnd)));
236      Parameters.Add(new ValueParameter<IntValue>("TestSamplesStart", new IntValue(testSamplesStart)));
237      Parameters.Add(new ValueParameter<IntValue>("TestSamplesEnd", new IntValue(testSamplesEnd)));
238      Parameters.Add(new ValueParameter<PercentValue>("ValidationPercentage", "The relative amount of the training samples that should be used as validation set.", new PercentValue(0.5)));
239
240      DatasetParameter.ReactOnValueToStringChangedAndValueItemImageChanged = false;
241      RegisterParameterEventHandlers();
242      RegisterParameterValueEventHandlers();
243    }
244
245    public override IDeepCloneable Clone(Cloner cloner) {
246      return new DataAnalysisProblemData(this, cloner);
247    }
248
249    [StorableHook(HookType.AfterDeserialization)]
250    private void AfterDeserialization() {
251      if (!Parameters.ContainsKey("ValidationPercentage"))
252        Parameters.Add(new ValueParameter<PercentValue>("ValidationPercentage", "The relative amount of the training samples that should be used as validation set.", new PercentValue(0.5)));
253
254      DatasetParameter.ReactOnValueToStringChangedAndValueItemImageChanged = false;
255      RegisterParameterEventHandlers();
256      RegisterParameterValueEventHandlers();
257    }
258
259    #region events
260    public event EventHandler ProblemDataChanged;
261    protected virtual void OnProblemDataChanged(EventArgs e) {
262      if (TrainingSamplesStart.Value < 0) TrainingSamplesStart.Value = 0;
263      else if (TestSamplesStart.Value < 0) TestSamplesStart.Value = 0;
264      else if (TrainingSamplesEnd.Value > Dataset.Rows - 1) TrainingSamplesEnd.Value = Dataset.Rows - 1;
265      else if (TestSamplesEnd.Value > Dataset.Rows - 1) TestSamplesEnd.Value = Dataset.Rows - 1;
266      else if (TrainingSamplesStart.Value > TrainingSamplesEnd.Value) TrainingSamplesStart.Value = TestSamplesEnd.Value;
267      else if (TestSamplesStart.Value > TestSamplesEnd.Value) TestSamplesStart.Value = TestSamplesEnd.Value;
268      else if (ValidationPercentage.Value < 0) ValidationPercentage.Value = 0;
269      else if (ValidationPercentage.Value > 1) ValidationPercentage.Value = 1;
270      else if (!TrainingIndizes.Any()) throw new ArgumentException("No training samples are available.");
271      else if (!suppressEvents) {
272        var listeners = ProblemDataChanged;
273        if (listeners != null) listeners(this, e);
274      }
275    }
276
277    private void RegisterParameterEventHandlers() {
278      DatasetParameter.ValueChanged += new EventHandler(DatasetParameter_ValueChanged);
279      InputVariablesParameter.ValueChanged += new EventHandler(InputVariablesParameter_ValueChanged);
280      TargetVariableParameter.ValueChanged += new EventHandler(TargetVariableParameter_ValueChanged);
281      TrainingSamplesStartParameter.ValueChanged += new EventHandler(TrainingSamplesStartParameter_ValueChanged);
282      TrainingSamplesEndParameter.ValueChanged += new EventHandler(TrainingSamplesEndParameter_ValueChanged);
283      TestSamplesStartParameter.ValueChanged += new EventHandler(TestSamplesStartParameter_ValueChanged);
284      TestSamplesEndParameter.ValueChanged += new EventHandler(TestSamplesEndParameter_ValueChanged);
285      ValidationPercentageParameter.ValueChanged += new EventHandler(ValidationPercentageParameter_ValueChanged);
286    }
287
288    private void RegisterParameterValueEventHandlers() {
289      RegisterInputVariablesEventHandlers();
290      if (TargetVariable != null) RegisterStringValueEventHandlers(TargetVariable);
291      RegisterValueTypeEventHandlers(TrainingSamplesStart);
292      RegisterValueTypeEventHandlers(TrainingSamplesEnd);
293      RegisterValueTypeEventHandlers(TestSamplesStart);
294      RegisterValueTypeEventHandlers(TestSamplesEnd);
295      RegisterValueTypeEventHandlers(ValidationPercentage);
296    }
297
298
299    #region parameter value changed event handlers
300    private void DatasetParameter_ValueChanged(object sender, EventArgs e) {
301      OnProblemDataChanged(EventArgs.Empty);
302    }
303    private void InputVariablesParameter_ValueChanged(object sender, EventArgs e) {
304      RegisterInputVariablesEventHandlers();
305      OnProblemDataChanged(EventArgs.Empty);
306    }
307    private void TargetVariableParameter_ValueChanged(object sender, EventArgs e) {
308      if (TargetVariable != null) {
309        RegisterStringValueEventHandlers(TargetVariable);
310        OnProblemDataChanged(EventArgs.Empty);
311      }
312    }
313    private void TrainingSamplesStartParameter_ValueChanged(object sender, EventArgs e) {
314      RegisterValueTypeEventHandlers(TrainingSamplesStart);
315      OnProblemDataChanged(EventArgs.Empty);
316    }
317    private void TrainingSamplesEndParameter_ValueChanged(object sender, EventArgs e) {
318      RegisterValueTypeEventHandlers(TrainingSamplesEnd);
319      OnProblemDataChanged(EventArgs.Empty);
320    }
321    private void TestSamplesStartParameter_ValueChanged(object sender, EventArgs e) {
322      RegisterValueTypeEventHandlers(TestSamplesStart);
323      OnProblemDataChanged(EventArgs.Empty);
324    }
325    private void TestSamplesEndParameter_ValueChanged(object sender, EventArgs e) {
326      RegisterValueTypeEventHandlers(TestSamplesEnd);
327      OnProblemDataChanged(EventArgs.Empty);
328    }
329    private void ValidationPercentageParameter_ValueChanged(object sender, EventArgs e) {
330      RegisterValueTypeEventHandlers(ValidationPercentage);
331      OnProblemDataChanged(EventArgs.Empty);
332    }
333    #endregion
334
335    private void RegisterInputVariablesEventHandlers() {
336      InputVariables.CollectionReset += new HeuristicLab.Collections.CollectionItemsChangedEventHandler<HeuristicLab.Collections.IndexedItem<StringValue>>(InputVariables_CollectionReset);
337      InputVariables.ItemsAdded += new HeuristicLab.Collections.CollectionItemsChangedEventHandler<HeuristicLab.Collections.IndexedItem<StringValue>>(InputVariables_ItemsAdded);
338      InputVariables.ItemsRemoved += new HeuristicLab.Collections.CollectionItemsChangedEventHandler<HeuristicLab.Collections.IndexedItem<StringValue>>(InputVariables_ItemsRemoved);
339      InputVariables.CheckedItemsChanged += new HeuristicLab.Collections.CollectionItemsChangedEventHandler<HeuristicLab.Collections.IndexedItem<StringValue>>(InputVariables_CheckedItemsChanged);
340      foreach (var item in InputVariables) {
341        item.ValueChanged += new EventHandler(InputVariable_ValueChanged);
342      }
343    }
344
345    private void DeregisterInputVariablesEventHandlers() {
346      InputVariables.CollectionReset -= new HeuristicLab.Collections.CollectionItemsChangedEventHandler<HeuristicLab.Collections.IndexedItem<StringValue>>(InputVariables_CollectionReset);
347      InputVariables.ItemsAdded -= new HeuristicLab.Collections.CollectionItemsChangedEventHandler<HeuristicLab.Collections.IndexedItem<StringValue>>(InputVariables_ItemsAdded);
348      InputVariables.ItemsRemoved -= new HeuristicLab.Collections.CollectionItemsChangedEventHandler<HeuristicLab.Collections.IndexedItem<StringValue>>(InputVariables_ItemsRemoved);
349      InputVariables.CheckedItemsChanged -= new HeuristicLab.Collections.CollectionItemsChangedEventHandler<HeuristicLab.Collections.IndexedItem<StringValue>>(InputVariables_CheckedItemsChanged);
350      foreach (var item in InputVariables) {
351        item.ValueChanged -= new EventHandler(InputVariable_ValueChanged);
352      }
353    }
354
355    private void InputVariables_CheckedItemsChanged(object sender, HeuristicLab.Collections.CollectionItemsChangedEventArgs<HeuristicLab.Collections.IndexedItem<StringValue>> e) {
356      OnProblemDataChanged(e);
357    }
358    private void InputVariables_ItemsRemoved(object sender, HeuristicLab.Collections.CollectionItemsChangedEventArgs<HeuristicLab.Collections.IndexedItem<StringValue>> e) {
359      foreach (var indexedItem in e.Items)
360        indexedItem.Value.ValueChanged -= new EventHandler(InputVariable_ValueChanged);
361      OnProblemDataChanged(e);
362    }
363    private void InputVariables_ItemsAdded(object sender, HeuristicLab.Collections.CollectionItemsChangedEventArgs<HeuristicLab.Collections.IndexedItem<StringValue>> e) {
364      foreach (var indexedItem in e.Items)
365        indexedItem.Value.ValueChanged += new EventHandler(InputVariable_ValueChanged);
366      OnProblemDataChanged(e);
367    }
368    private void InputVariables_CollectionReset(object sender, HeuristicLab.Collections.CollectionItemsChangedEventArgs<HeuristicLab.Collections.IndexedItem<StringValue>> e) {
369      foreach (var indexedItem in e.OldItems)
370        indexedItem.Value.ValueChanged -= new EventHandler(InputVariable_ValueChanged);
371      OnProblemDataChanged(e);
372    }
373    private void InputVariable_ValueChanged(object sender, EventArgs e) {
374      OnProblemDataChanged(e);
375    }
376
377    #region helper
378    private void RegisterValueTypeEventHandlers<T>(ValueTypeValue<T> value) where T : struct {
379      value.ValueChanged += new EventHandler(value_ValueChanged);
380    }
381    private void DeregisterValueTypeEventHandlers<T>(ValueTypeValue<T> value) where T : struct {
382      value.ValueChanged -= new EventHandler(value_ValueChanged);
383    }
384    private void RegisterStringValueEventHandlers(StringValue value) {
385      value.ValueChanged += new EventHandler(value_ValueChanged);
386    }
387    private void DeregisterStringValueEventHandlers(StringValue value) {
388      value.ValueChanged -= new EventHandler(value_ValueChanged);
389    }
390
391    private void value_ValueChanged(object sender, EventArgs e) {
392      OnProblemDataChanged(e);
393    }
394    #endregion
395    #endregion
396
397    public virtual void ImportFromFile(string fileName) {
398      var csvFileParser = new TableFileParser();
399      csvFileParser.Parse(fileName);
400      suppressEvents = true;
401      Name = "Data imported from " + Path.GetFileName(fileName);
402      Dataset = new Dataset(csvFileParser.VariableNames, csvFileParser.Values);
403      Dataset.Name = Path.GetFileName(fileName);
404      var variableNames = Dataset.VariableNames.Select(x => new StringValue(x).AsReadOnly()).ToList();
405      ((ConstrainedValueParameter<StringValue>)TargetVariableParameter).ValidValues.Clear();
406      foreach (var variableName in variableNames)
407        ((ConstrainedValueParameter<StringValue>)TargetVariableParameter).ValidValues.Add(variableName);
408      TargetVariable = variableNames.First();
409      InputVariables = new CheckedItemList<StringValue>(variableNames).AsReadOnly();
410      InputVariables.SetItemCheckedState(variableNames.First(), false);
411      int middle = (int)(csvFileParser.Rows * 0.5);
412      TrainingSamplesEnd = new IntValue(middle);
413      TrainingSamplesStart = new IntValue(0);
414      TestSamplesEnd = new IntValue(csvFileParser.Rows);
415      TestSamplesStart = new IntValue(middle);
416      suppressEvents = false;
417      OnProblemDataChanged(EventArgs.Empty);
418    }
419  }
420}
Note: See TracBrowser for help on using the repository browser.