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
|
/**
* Solve classification problem by random forest.
*/
#ifndef CLASSIFICATION_H
#define CLASSIFICATION_H
#include <set>
#include <map>
#include "trainer.h"
// Bug - Paul (changed labelT to be a parameter)
template<class dataT, class labelT, class ClassifierT>
class Classification
{
public:
// typedef int labelT;
typedef Histogram<dataT, labelT> HistStatisticsT;
typedef ClassificationContext<ClassifierT, dataT, labelT> ClassificationContextT;
typedef DecisionForest<HistStatisticsT, ClassifierT, dataT> DecisionForestT;
typedef Trainer<ClassifierT, HistStatisticsT, dataT, labelT> TrainerT;
// Bug - Paul (changed int to labelT)
typedef MLData<dataT, labelT> TrainingDataT;
typedef MLData<dataT, HistStatisticsT*> TestingDataT;
typedef Vector<Vector<HistStatisticsT*> > TestingResultT;
typedef Matrix<double> SoftPredictionT;
typedef Vector<int> HardPredictionT;
bool ValidData(TrainingDataT& trainingData,
std::map<index_t, labelT>& mapping)
{
std::set<int> labelSet(trainingData.label.Begin(),
trainingData.label.End());
std::set<int>::iterator setIter;
std::map<int, index_t> backMapping;
mapping.clear();
index_t i = 0;
bool valid = true;
for (setIter = labelSet.begin(); setIter != labelSet.end(); ++setIter, ++i)
{
backMapping.insert(std::map<int, index_t>::value_type(*setIter, i));
mapping.insert(std::map<index_t, int>::value_type(i, *setIter));
if (*setIter != i)
{
valid = false;
}
}
if (valid == false)
{
std::map<int, index_t>::iterator backMappingIter;
for (index_t i = 0; i < trainingData.Size(); ++i)
{
backMappingIter = backMapping.find(trainingData.label[i]);
if (backMappingIter == backMapping.end())
{
throw std::runtime_error("Classificaiton: ValidData error");
}
else
{
trainingData.label[i] = backMappingIter->second;
}
}
}
return valid;
}
// Bug: paul - int should be labelT
void Learning(TrainingParameters& trainingParameters,
TrainingDataT& trainingData,
DecisionForestT& forest,
bool& validLabel,
std::map<index_t, labelT>& mapping)
{
validLabel = ValidData(trainingData, mapping);
size_t classNum = mapping.size();
if ( !trainingParameters.weights.empty() )
{
if (classNum != trainingParameters.weights.size())
{
throw std::runtime_error("Traing parameter weights size doesn't match class size in data\n");
}
}
Random random;
ClassificationContextT classificationTC(trainingData.Dimension(), classNum);
TrainerT trainer(trainingData, trainingParameters, classificationTC, random);
trainer.Training(forest);
}
void Predicting(DecisionForestT& forest,
TestingDataT& testingData,
bool& validLabel, std::map<index_t, int>& mapping,
SoftPredictionT& softPrediction,
HardPredictionT& hardPrediction)
{
TestingResultT testingResult;
forest.Apply(testingData, testingResult);
size_t classNum = mapping.size();
size_t treeNum = testingResult.Size();
size_t dataNum = testingResult[0].Size();
std::vector<double> max(dataNum, 0);
#pragma omp parallel for
for (index_t i = 0; i < dataNum; ++i)
{
for (index_t j = 0; j < classNum; ++j)
{
for (index_t k = 0; k < treeNum; ++k)
{
if (k == 0)
{
softPrediction[i][j] = testingResult[k][i]->prob_[j];
}
else
{
softPrediction[i][j] = softPrediction[i][j] + testingResult[k][i]->prob_[j];
}
}
softPrediction[i][j] = softPrediction[i][j] / treeNum;
if (softPrediction[i][j] > max[i])
{
max[i] = softPrediction[i][j];
hardPrediction[i] = j;
}
}
if (validLabel == false)
{
// Bug: typename missing
typename std::map<index_t, labelT>::iterator mapIter = mapping.find(hardPrediction[i]);
if (mapIter == mapping.end())
{
throw std::runtime_error("Classificaiton: Predicting mapping error");
}
else
{
hardPrediction[i] = mapIter->second;
}
}
}
}
void Run(TrainingParameters& trainingParameters,
TrainingDataT& trainingData,
TestingDataT& testingData,
DecisionForestT& forest,
std::map<index_t, int>& mapping,
SoftPredictionT& softPrediction,
HardPredictionT& hardPrediction)
{
bool validLabel = true;
MPTimer timer;
std::cerr << "training begin\n";
timer.Start();
Learning(trainingParameters, trainingData, forest, validLabel, mapping);
std::cerr << "training finished, spending " << timer.StopAndSpendSecond() << " secs\n";
std::cerr << "testing begin\n";
timer.Start();
Predicting(forest, testingData, validLabel, mapping, softPrediction, hardPrediction);
std::cerr << "testing finished, spending " << timer.StopAndSpendSecond() << " secs\n";
}
};
#endif // CLASSIFICATION_H
|
