Free cookie consent management tool by TermsFeed Policy Generator

source: trunk/sources/HeuristicLab.DataPreprocessing/3.4/Logic/StatisticsLogic.cs @ 13935

Last change on this file since 13935 was 13935, checked in by mkommend, 8 years ago

#2616: Added default value for min, max and most common value operations for data preprocessing.

File size: 8.4 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2015 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.Common;
26
27namespace HeuristicLab.DataPreprocessing {
28  public class StatisticsLogic {
29
30    private readonly ITransactionalPreprocessingData preprocessingData;
31    private readonly SearchLogic searchLogic;
32
33    public StatisticsLogic(ITransactionalPreprocessingData thePreprocessingData, SearchLogic theSearchLogic) {
34      preprocessingData = thePreprocessingData;
35      searchLogic = theSearchLogic;
36    }
37
38    public int GetColumnCount() {
39      return searchLogic.Columns;
40    }
41
42    public int GetRowCount() {
43      return searchLogic.Rows;
44    }
45
46    public int GetNumericColumnCount() {
47      int count = 0;
48
49      for (int i = 0; i < searchLogic.Columns; ++i) {
50        if (preprocessingData.VariableHasType<double>(i)) {
51          ++count;
52        }
53      }
54      return count;
55    }
56
57    public int GetNominalColumnCount() {
58      return searchLogic.Columns - GetNumericColumnCount();
59    }
60
61    public int GetMissingValueCount() {
62      int count = 0;
63      for (int i = 0; i < searchLogic.Columns; ++i) {
64        count += GetMissingValueCount(i);
65      }
66      return count;
67    }
68
69    public int GetMissingValueCount(int columnIndex) {
70      return searchLogic.GetMissingValueIndices(columnIndex).Count();
71    }
72
73    public T GetMin<T>(int columnIndex, T defaultValue, bool considerSelection = false) where T : IComparable<T> {
74      var min = defaultValue;
75      if (preprocessingData.VariableHasType<T>(columnIndex)) {
76        var values = GetValuesWithoutNaN<T>(columnIndex, considerSelection);
77        if (values.Any()) {
78          min = values.Min();
79        }
80      }
81      return min;
82    }
83
84    public T GetMax<T>(int columnIndex, T defaultValue, bool considerSelection = false) where T : IComparable<T> {
85      var max = defaultValue;
86      if (preprocessingData.VariableHasType<T>(columnIndex)) {
87        var values = GetValuesWithoutNaN<T>(columnIndex, considerSelection);
88        if (values.Any()) {
89          max = values.Max();
90        }
91      }
92      return max;
93    }
94
95    public double GetMedian(int columnIndex, bool considerSelection = false) {
96      double median = double.NaN;
97      if (preprocessingData.VariableHasType<double>(columnIndex)) {
98        var values = GetValuesWithoutNaN<double>(columnIndex, considerSelection);
99        if (values.Any()) {
100          median = values.Median();
101        }
102      }
103      return median;
104    }
105
106    public double GetAverage(int columnIndex, bool considerSelection = false) {
107      double avg = double.NaN;
108      if (preprocessingData.VariableHasType<double>(columnIndex)) {
109        var values = GetValuesWithoutNaN<double>(columnIndex, considerSelection);
110        if (values.Any()) {
111          avg = values.Average();
112        }
113      }
114      return avg;
115    }
116
117    public DateTime GetMedianDateTime(int columnIndex, bool considerSelection = false) {
118      DateTime median = new DateTime();
119      if (preprocessingData.VariableHasType<DateTime>(columnIndex)) {
120        median = GetSecondsAsDateTime(GetDateTimeAsSeconds(columnIndex, considerSelection).Median());
121      }
122      return median;
123    }
124
125    public DateTime GetAverageDateTime(int columnIndex, bool considerSelection = false) {
126      DateTime avg = new DateTime();
127      if (preprocessingData.VariableHasType<DateTime>(columnIndex)) {
128        avg = GetSecondsAsDateTime(GetDateTimeAsSeconds(columnIndex, considerSelection).Average());
129      }
130      return avg;
131    }
132
133    public T GetMostCommonValue<T>(int columnIndex, T defaultValue, bool considerSelection = false) {
134      var values = GetValuesWithoutNaN<T>(columnIndex, considerSelection);
135      if (!values.Any())
136        return defaultValue;
137      return values.GroupBy(x => x)
138                              .OrderByDescending(g => g.Count())
139                              .Select(g => g.Key)
140                              .First();
141    }
142
143
144    public double GetStandardDeviation(int columnIndex) {
145      double stdDev = double.NaN;
146      if (preprocessingData.VariableHasType<double>(columnIndex)) {
147        stdDev = GetValuesWithoutNaN<double>(columnIndex).StandardDeviation();
148      } else if (preprocessingData.VariableHasType<DateTime>(columnIndex)) {
149        stdDev = GetDateTimeAsSeconds(columnIndex).StandardDeviation();
150      }
151      return stdDev;
152    }
153
154    public double GetVariance(int columnIndex) {
155      double variance = double.NaN;
156      if (preprocessingData.VariableHasType<double>(columnIndex)) {
157        variance = GetValuesWithoutNaN<double>(columnIndex).Variance();
158      } else if (preprocessingData.VariableHasType<DateTime>(columnIndex)) {
159        variance = GetDateTimeAsSeconds(columnIndex).Variance();
160      }
161      return variance;
162    }
163
164    public double GetOneQuarterPercentile(int columnIndex) {
165      double percentile = double.NaN;
166      if (preprocessingData.VariableHasType<double>(columnIndex)) {
167        percentile = GetValuesWithoutNaN<double>(columnIndex).DefaultIfEmpty(double.NaN).Quantile(0.25);
168      } else if (preprocessingData.VariableHasType<DateTime>(columnIndex)) {
169        percentile = GetDateTimeAsSeconds(columnIndex).DefaultIfEmpty(double.NaN).Quantile(0.25);
170      }
171      return percentile;
172    }
173
174    public double GetThreeQuarterPercentile(int columnIndex) {
175      double percentile = double.NaN;
176      if (preprocessingData.VariableHasType<double>(columnIndex)) {
177        percentile = GetValuesWithoutNaN<double>(columnIndex).DefaultIfEmpty(double.NaN).Quantile(0.75);
178      } else if (preprocessingData.VariableHasType<DateTime>(columnIndex)) {
179        percentile = GetDateTimeAsSeconds(columnIndex).DefaultIfEmpty(double.NaN).Quantile(0.75);
180      }
181      return percentile;
182    }
183
184    public int GetDifferentValuesCount<T>(int columnIndex) {
185      return preprocessingData.GetValues<T>(columnIndex).GroupBy(x => x).Count();
186    }
187
188    public int GetRowMissingValueCount(int rowIndex) {
189      int count = 0;
190      for (int i = 0; i < preprocessingData.Columns; ++i) {
191        if (searchLogic.IsMissingValue(i, rowIndex)) {
192          ++count;
193        }
194      }
195      return count;
196    }
197
198    public string GetVariableName(int columnIndex) {
199      return preprocessingData.GetVariableName(columnIndex);
200    }
201
202    public bool VariableHasType<T>(int columnIndex) {
203      return preprocessingData.VariableHasType<T>(columnIndex);
204    }
205
206    public string GetColumnTypeAsString(int columnIndex) {
207      if (preprocessingData.VariableHasType<double>(columnIndex)) {
208        return "double";
209      } else if (preprocessingData.VariableHasType<string>(columnIndex)) {
210        return "string";
211      } else if (preprocessingData.VariableHasType<DateTime>(columnIndex)) {
212        return "DateTime";
213      }
214      return "Unknown Type";
215    }
216
217    private IEnumerable<double> GetDateTimeAsSeconds(int columnIndex, bool considerSelection = false) {
218      return GetValuesWithoutNaN<DateTime>(columnIndex, considerSelection).Select(x => (double)x.Ticks / TimeSpan.TicksPerSecond);
219    }
220
221    private IEnumerable<T> GetValuesWithoutNaN<T>(int columnIndex, bool considerSelection = false) {
222      return searchLogic.GetValuesWithoutNaN<T>(columnIndex, considerSelection);
223    }
224
225    private DateTime GetSecondsAsDateTime(double seconds) {
226      DateTime dateTime = new DateTime();
227      return dateTime.AddSeconds(seconds);
228    }
229
230    public event DataPreprocessingChangedEventHandler Changed {
231      add { preprocessingData.Changed += value; }
232      remove { preprocessingData.Changed -= value; }
233    }
234  }
235}
Note: See TracBrowser for help on using the repository browser.