Logo ROOT   6.18/05
Reference Guide
MethodPyRandomForest.h
Go to the documentation of this file.
1// @(#)root/tmva/pymva $Id$
2// Authors: Omar Zapata, Lorenzo Moneta, Sergei Gleyzer 2015
3
4/**********************************************************************************
5 * Project: TMVA - a Root-integrated toolkit for multivariate data analysis *
6 * Package: TMVA *
7 * Class : MethodPyRandomForest *
8 * Web : http://oproject.org *
9 * *
10 * Description: *
11 * scikit-learn Package RandomForestClassifier method based on python *
12 * *
13 **********************************************************************************/
14
15#ifndef ROOT_TMVA_MethodPyRandomForest
16#define ROOT_TMVA_MethodPyRandomForest
17
18//////////////////////////////////////////////////////////////////////////
19// //
20// MethodPyRandomForest //
21// //
22//////////////////////////////////////////////////////////////////////////
23
24#include "TMVA/PyMethodBase.h"
25
26namespace TMVA {
27
28 class Factory; // DSMTEST
29 class Reader; // DSMTEST
30 class DataSetManager; // DSMTEST
31 class Types;
33
34 public :
35 // constructors
36 MethodPyRandomForest(const TString &jobName,
37 const TString &methodTitle,
38 DataSetInfo &theData,
39 const TString &theOption = "");
40
42 const TString &theWeightFile);
43
45 void Train();
46
47 // options treatment
48 void Init();
49 void DeclareOptions();
50 void ProcessOptions();
51
52 // create ranking
53 const Ranking *CreateRanking();
54
55 Bool_t HasAnalysisType(Types::EAnalysisType type, UInt_t numberClasses, UInt_t numberTargets);
56
57 // performs classifier testing
58 virtual void TestClassification();
59
60 // Get class probabilities of given event
61 Double_t GetMvaValue(Double_t *errLower = 0, Double_t *errUpper = 0);
62 std::vector<Double_t> GetMvaValues(Long64_t firstEvt = 0, Long64_t lastEvt = -1, Bool_t logProgress = false);
63 std::vector<Float_t>& GetMulticlassValues();
64
66 // the actual "weights"
67 virtual void AddWeightsXMLTo(void * /* parent */) const {} // = 0;
68 virtual void ReadWeightsFromXML(void * /* wghtnode */) {} // = 0;
69 virtual void ReadWeightsFromStream(std::istream &) {} //= 0; // backward compatibility
70
71 void ReadModelFromFile();
72
73 private :
75 friend class Factory; // DSMTEST
76 friend class Reader; // DSMTEST
77
78 protected:
79 std::vector<Double_t> mvaValues;
80 std::vector<Float_t> classValues;
81
82 UInt_t fNvars; // number of variables
83 UInt_t fNoutputs; // number of outputs
84 TString fFilenameClassifier; // Path to serialized classifier (default in `weights` folder)
85
86 // RandomForest options
87
89 Int_t fNestimators; //integer, optional (default=10)
90 //The number of trees in the forest.
91
93 TString fCriterion; //string, optional (default="gini")
94 //The function to measure the quality of a split. Supported criteria are
95 //"gini" for the Gini impurity and "entropy" for the information gain.
96 //Note: this parameter is tree-specific.
97
99 TString fMaxDepth; //integer or None, optional (default=None)
100 //The maximum depth of the tree. If None, then nodes are expanded until
101 //all leaves are pure or until all leaves contain less than `fMinSamplesSplit`.
102
104 Int_t fMinSamplesSplit; //integer, optional (default=2)
105 //The minimum number of samples required to split an internal node.
106
108 Int_t fMinSamplesLeaf; //integer, optional (default=1)
109 //The minimum number of samples in newly created leaves. A split is
110 //discarded if after the split, one of the leaves would contain less then
111 //``min_samples_leaf`` samples.
112 //Note: this parameter is tree-specific.
113
115 Double_t fMinWeightFractionLeaf; //float, optional (default=0.)
116 //The minimum weighted fraction of the input samples required to be at a
117 //leaf node.
118 //Note: this parameter is tree-specific.
119
121 TString fMaxFeatures; //int, float, string or None, optional (default="auto")
122 //The number of features to consider when looking for the best split:
123 //- If int, then consider `max_features` features at each split.
124 //- If float, then `max_features` is a percentage and
125 //`int(max_features * n_features)` features are considered at each split.
126 //- If "auto", then `max_features=sqrt(n_features)`.
127 //- If "sqrt", then `max_features=sqrt(n_features)`.
128 //- If "log2", then `max_features=log2(n_features)`.
129 //- If None, then `max_features=n_features`.
130 // Note: the search for a split does not stop until at least one
131 // valid partition of the node samples is found, even if it requires to
132 // effectively inspect more than ``max_features`` features.
133 // Note: this parameter is tree-specific.
134
136 TString fMaxLeafNodes; //int or None, optional (default=None)
137 //Grow trees with ``max_leaf_nodes`` in best-first fashion.
138 //Best nodes are defined as relative reduction in impurity.
139 //If None then unlimited number of leaf nodes.
140 //If not None then ``max_depth`` will be ignored.
141
143 Bool_t fBootstrap; //boolean, optional (default=True)
144 //Whether bootstrap samples are used when building trees.
145
147 Bool_t fOobScore; //Whether to use out-of-bag samples to estimate
148 //the generalization error.
149
151 Int_t fNjobs; // integer, optional (default=1)
152 //The number of jobs to run in parallel for both `fit` and `predict`.
153 //If -1, then the number of jobs is set to the number of cores.
154
156 TString fRandomState; //int, RandomState instance or None, optional (default=None)
157 //If int, random_state is the seed used by the random number generator;
158 //If RandomState instance, random_state is the random number generator;
159 //If None, the random number generator is the RandomState instance used
160 //by `np.random`.
161
163 Int_t fVerbose; //Controls the verbosity of the tree building process.
164
166 Bool_t fWarmStart; //bool, optional (default=False)
167 //When set to ``True``, reuse the solution of the previous call to fit
168 //and add more estimators to the ensemble, otherwise, just fit a whole
169 //new forest.
170
172 TString fClassWeight; //dict, list of dicts, "auto", "subsample" or None, optional
173 //Weights associated with classes in the form ``{class_label: weight}``.
174 //If not given, all classes are supposed to have weight one. For
175 //multi-output problems, a list of dicts can be provided in the same
176 //order as the columns of y.
177 //The "auto" mode uses the values of y to automatically adjust
178 //weights inversely proportional to class frequencies in the input data.
179 //The "subsample" mode is the same as "auto" except that weights are
180 //computed based on the bootstrap sample for every tree grown.
181 //For multi-output, the weights of each column of y will be multiplied.
182 //Note that these weights will be multiplied with sample_weight (passed
183 //through the fit method) if sample_weight is specified.
184
185 // get help message text
186 void GetHelpMessage() const;
187
189 };
190
191} // namespace TMVA
192
193#endif // ROOT_TMVA_MethodPyRandomForest
int Int_t
Definition: RtypesCore.h:41
unsigned int UInt_t
Definition: RtypesCore.h:42
bool Bool_t
Definition: RtypesCore.h:59
double Double_t
Definition: RtypesCore.h:55
long long Long64_t
Definition: RtypesCore.h:69
#define ClassDef(name, id)
Definition: Rtypes.h:326
int type
Definition: TGX11.cxx:120
_object PyObject
Definition: TPyArg.h:20
Class that contains all the data information.
Definition: DataSetInfo.h:60
Class that contains all the data information.
This is the main MVA steering class.
Definition: Factory.h:81
virtual void ReadWeightsFromStream(std::istream &)=0
MethodPyRandomForest(const TString &jobName, const TString &methodTitle, DataSetInfo &theData, const TString &theOption="")
std::vector< Float_t > & GetMulticlassValues()
Bool_t HasAnalysisType(Types::EAnalysisType type, UInt_t numberClasses, UInt_t numberTargets)
std::vector< Double_t > GetMvaValues(Long64_t firstEvt=0, Long64_t lastEvt=-1, Bool_t logProgress=false)
get all the MVA values for the events of the current Data type
std::vector< Float_t > classValues
virtual void AddWeightsXMLTo(void *) const
std::vector< Double_t > mvaValues
virtual void ReadWeightsFromStream(std::istream &)
virtual void TestClassification()
initialization
virtual void ReadWeightsFromXML(void *)
Double_t GetMvaValue(Double_t *errLower=0, Double_t *errUpper=0)
Ranking for variables in method (implementation)
Definition: Ranking.h:48
The Reader class serves to use the MVAs in a specific analysis context.
Definition: Reader.h:63
EAnalysisType
Definition: Types.h:127
Basic string class.
Definition: TString.h:131
create variable transformations