SHOGUN  4.1.0
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Modules Pages
RandomSearchModelSelection.cpp
Go to the documentation of this file.
1 /*
2  * This program is free software; you can redistribute it and/or modify
3  * it under the terms of the GNU General Public License as published by
4  * the Free Software Foundation; either version 3 of the License, or
5  * (at your option) any later version.
6  *
7  * Copyright (C) 2011 Heiko Strathmann
8  * Copyright (C) 2012 Sergey Lisitsyn
9  */
10 
16 #include <shogun/machine/Machine.h>
17 
18 using namespace shogun;
19 
21 {
22  set_ratio(0.5);
23 }
24 
26  CMachineEvaluation* machine_eval,
27  CModelSelectionParameters* model_parameters, float64_t ratio)
28  : CModelSelection(machine_eval, model_parameters)
29 {
30  set_ratio(ratio);
31 }
32 
34 {
35 }
36 
38 {
39  if (print_state)
40  SG_PRINT("Generating parameter combinations\n")
41 
42  /* Retrieve all possible parameter combinations */
43  CDynamicObjectArray* all_combinations=
45 
46  int32_t n_all_combinations=all_combinations->get_num_elements();
47  SGVector<index_t> combinations_indices=CStatistics::sample_indices(n_all_combinations*m_ratio, n_all_combinations);
48 
49  CDynamicObjectArray* combinations=new CDynamicObjectArray();
50 
51  for (int32_t i=0; i<combinations_indices.vlen; i++)
52  combinations->append_element(all_combinations->get_element(i));
53 
55 
56  CParameterCombination* best_combination=NULL;
58  {
59  if (print_state) SG_PRINT("Direction is maximize\n")
60  best_result->mean=CMath::ALMOST_NEG_INFTY;
61  }
62  else
63  {
64  if (print_state) SG_PRINT("Direction is minimize\n")
65  best_result->mean=CMath::ALMOST_INFTY;
66  }
67 
68  /* underlying learning machine */
70 
71  /* apply all combinations and search for best one */
72  for (index_t i=0; i<combinations->get_num_elements(); ++i)
73  {
74  CParameterCombination* current_combination=(CParameterCombination*)
75  combinations->get_element(i);
76 
77  /* eventually print */
78  if (print_state)
79  {
80  SG_PRINT("trying combination:\n")
81  current_combination->print_tree();
82  }
83 
84  current_combination->apply_to_modsel_parameter(
86 
87  /* note that this may implicitly lock and unlockthe machine */
88  CCrossValidationResult* result =
90 
91  if (result->get_result_type() != CROSSVALIDATION_RESULT)
92  SG_ERROR("Evaluation result is not of type CCrossValidationResult!")
93 
94  if (print_state)
95  result->print_result();
96 
97  /* check if current result is better, delete old combinations */
99  {
100  if (result->mean>best_result->mean)
101  {
102  if (best_combination)
103  SG_UNREF(best_combination);
104 
105  best_combination=(CParameterCombination*)
106  combinations->get_element(i);
107 
108  SG_REF(result);
109  SG_UNREF(best_result);
110  best_result=result;
111  }
112  else
113  {
115  combinations->get_element(i);
116  SG_UNREF(combination);
117  }
118  }
119  else
120  {
121  if (result->mean<best_result->mean)
122  {
123  if (best_combination)
124  SG_UNREF(best_combination);
125 
126  best_combination=(CParameterCombination*)
127  combinations->get_element(i);
128 
129  SG_REF(result);
130  SG_UNREF(best_result);
131  best_result=result;
132  }
133  else
134  {
136  combinations->get_element(i);
137  SG_UNREF(combination);
138  }
139  }
140 
141  SG_UNREF(result);
142  SG_UNREF(current_combination);
143  }
144 
145  SG_UNREF(best_result);
146  SG_UNREF(machine);
147  SG_UNREF(combinations);
148 
149  return best_combination;
150 }
static SGVector< int32_t > sample_indices(int32_t sample_size, int32_t N)
virtual CParameterCombination * select_model(bool print_state=false)
virtual EEvaluationResultType get_result_type() const
int32_t index_t
Definition: common.h:62
void apply_to_modsel_parameter(Parameter *parameter) const
#define SG_ERROR(...)
Definition: SGIO.h:129
void print_tree(int prefix_num=0) const
type to encapsulate the results of an evaluation run. May contain confidence interval (if conf_int_al...
#define SG_REF(x)
Definition: SGObject.h:51
A generic learning machine interface.
Definition: Machine.h:143
Class to select parameters and their ranges for model selection. The structure is organized as a tree...
virtual CEvaluationResult * evaluate()=0
static const float64_t ALMOST_NEG_INFTY
almost neg (log) infinity
Definition: Math.h:2052
Abstract base class for model selection.
CModelSelectionParameters * m_model_parameters
CMachineEvaluation * m_machine_eval
index_t vlen
Definition: SGVector.h:494
#define SG_PRINT(...)
Definition: SGIO.h:137
CMachine * get_machine() const
double float64_t
Definition: common.h:50
Machine Evaluation is an abstract class that evaluates a machine according to some criterion...
Dynamic array class for CSGObject pointers that creates an array that can be used like a list or an a...
Class that holds ONE combination of parameters for a learning machine. The structure is organized as ...
EEvaluationDirection get_evaluation_direction()
Parameter * m_model_selection_parameters
Definition: SGObject.h:508
#define SG_UNREF(x)
Definition: SGObject.h:52
all of classes and functions are contained in the shogun namespace
Definition: class_list.h:18
CDynamicObjectArray * get_combinations(index_t prefix_num=1)
static const float64_t ALMOST_INFTY
Definition: Math.h:2049
CSGObject * get_element(int32_t index) const
bool append_element(CSGObject *e)

SHOGUN Machine Learning Toolbox - Documentation