-
Notifications
You must be signed in to change notification settings - Fork 0
/
DData.cpp
260 lines (179 loc) · 6.27 KB
/
DData.cpp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
#include "DData.h"
void DData::loadFromCSV(const std::string& _fileName, unsigned int maxSamples)
{
std::ifstream ifile(_fileName);
std::string line;
std::string token;
unsigned int featureIndex = 0;
//ifile >> line;
std::getline(ifile, line);
std::istringstream iss(line);
while (std::getline(iss, token, ';'))
{
colNames.push_back(token);
featureIndices.push_back(featureIndex++);
}
//pop two: one for id, one for class
featureIndices.pop_back();
featureIndices.pop_back();
line.clear();
iss.clear();
while (samples.size() < maxSamples && ifile >> line)
{
DSample sample;
iss.str(line);
iss >> sample;
samples.push_back(sample);
line.clear();
iss.clear();
}
ifile.close();
isLoaded = !samples.empty();
}
void DData::enumerateUnordered()
{
if (!isLoaded)
return;
unsigned int featureIndex = 0;
//getFeatureCount() + 1, because we want to enumerate the target class at the back of the vector
unsigned int featureCount = samples[featureIndex].getFeatureCount() + 1;
std::vector<DSample>::iterator it = samples.begin();
for (featureIndex; featureIndex < featureCount; featureIndex++)
{
if ((*it)[featureIndex].isOrdered() && featureIndex < (*it).getFeatureCount())
continue;
double currentEnumeration = 0.0;
std::string currentLabel;
std::unordered_map<std::string, double> toNumeric;
for (it; it != samples.end(); it++)
{
currentLabel = (*it)[featureIndex].getStringValue();
if (toNumeric.find(currentLabel) != toNumeric.end())
{
(*it).getFeatureWriteAcessAt(featureIndex).setNumericValue(toNumeric[currentLabel]);
}
else
{
(*it).getFeatureWriteAcessAt(featureIndex).setNumericValue(currentEnumeration);
toNumeric.insert(std::make_pair(currentLabel, currentEnumeration++));
}
}
it = samples.begin();
}
}
DData::DData(const std::string& _fileName, unsigned int _maxSamples)
{
fileName = _fileName;
loadFromCSV(_fileName, _maxSamples);
enumerateUnordered();
}
const std::string& DData::getFileName() const
{
return fileName;
}
const std::vector<DSample>& DData::getSamples() const
{
return samples;
}
const std::vector<std::string>& DData::getColNames() const
{
return colNames;
}
const std::vector<unsigned int>& DData::getFeatureIndices() const
{
return featureIndices;
}
int DData::getSampleSize() const
{
return samples.size();
}
const DSample& DData::operator[](unsigned int index)const
{
if (!isLoaded)
{
std::cerr << "Data was not loaded into DData object!";
exit(1);
}
if (index > samples.size() - 1 || index < 0 || samples.empty())
{
std::cerr << "Out of bounds index in DData.samples array!";
exit(1);
}
return samples[index];
}
void DData::addSample(const DSample& sample)
{
if (!isLoaded)
return;
samples.push_back(sample);
std::ofstream ofile(fileName, std::ios_base::app);
ofile << sample;
ofile.close();
}
void DData::saveInside(const std::string& outputFileName) const
{
if (!isLoaded)
return;
std::ofstream ofile(outputFileName);
std::vector<std::string>::const_iterator it1;
for (it1 = colNames.begin(); it1 != colNames.end() - 1; it1++)
ofile << *it1 << ';';
ofile << *it1 << '\n';
std::vector<DSample>::const_iterator it2;
for (it2 = samples.begin(); it2 != samples.end(); it2++)
ofile << *it2;
ofile.close();
}
void DData::generateFeatureIndices(std::vector<unsigned int>& randomFeatureIndices, std::function<unsigned int(unsigned int)> featureFunc) const
{
if (samples.empty() || featureIndices.empty() || !isLoaded)
return;
unsigned int originalFeatureCount = featureIndices.size();
unsigned int mutatedFeatureCount = featureFunc(originalFeatureCount);
//make sure the vector is clean
randomFeatureIndices.clear();
//fill vector with original indices
randomFeatureIndices.assign(featureIndices.begin(), featureIndices.end());
//if mutated =/= original pick mutatedFeaturCount indices randomly, do the shuffle
if (mutatedFeatureCount < originalFeatureCount)
{
//randomly pick and place mutatedFeatureCount elements at the front of randomFeatureIndices
//*automatically stops shuffling after mutatedFeatureCount random elements are placed
random_unique(randomFeatureIndices.begin(), randomFeatureIndices.end(), mutatedFeatureCount);
//remove and free the memory for the indices that we dont need (which we did not pick)
randomFeatureIndices.erase(randomFeatureIndices.begin() + mutatedFeatureCount, randomFeatureIndices.end());
}
}
void DData::generateSampleIndices(std::vector<unsigned int>& sampleIndices, std::vector<double>& sampleWeights, bool bootstrappigAllowed) const
{
if (!isLoaded || samples.empty())
return;
sampleIndices.clear();
sampleWeights.clear();
if (bootstrappigAllowed)
{
std::random_device randomDevice;
std::mt19937 randomGenerator(randomDevice());
std::uniform_int_distribution<> uniformDistirbution(0, samples.size() - 1);
std::set<unsigned int> uniqueSampleIndices;
sampleWeights = std::vector<double>(samples.size());
unsigned int generatedIndex = 0;
for (std::vector<DSample>::const_iterator it = samples.begin(); it != samples.end(); it++)
{
generatedIndex = uniformDistirbution(randomGenerator);
uniqueSampleIndices.insert(generatedIndex);
sampleWeights[generatedIndex]++;
}
sampleIndices.assign(uniqueSampleIndices.begin(), uniqueSampleIndices.end());
}
else
{
unsigned int sampleIndex = 0;
const double sampleWeight = 1.0;
for (std::vector<DSample>::const_iterator it = samples.begin(); it != samples.end(); it++)
{
sampleIndices.push_back(sampleIndex++);
sampleWeights.push_back(sampleWeight);
}
}
}