Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #2913 from yorkerlin/develop
support sparse penalty (eg, L1)
- Loading branch information
Showing
14 changed files
with
850 additions
and
324 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,146 @@ | ||
| /* | ||
| * Copyright (c) The Shogun Machine Learning Toolbox | ||
| * Written (w) 2015 Wu Lin | ||
| * All rights reserved. | ||
| * | ||
| * Redistribution and use in source and binary forms, with or without | ||
| * modification, are permitted provided that the following conditions are met: | ||
| * | ||
| * 1. Redistributions of source code must retain the above copyright notice, this | ||
| * list of conditions and the following disclaimer. | ||
| * 2. Redistributions in binary form must reproduce the above copyright notice, | ||
| * this list of conditions and the following disclaimer in the documentation | ||
| * and/or other materials provided with the distribution. | ||
| * | ||
| * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND | ||
| * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED | ||
| * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE | ||
| * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR | ||
| * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES | ||
| * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; | ||
| * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND | ||
| * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT | ||
| * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS | ||
| * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | ||
| * | ||
| * The views and conclusions contained in the software and documentation are those | ||
| * of the authors and should not be interpreted as representing official policies, | ||
| * either expressed or implied, of the Shogun Development Team. | ||
| * | ||
| */ | ||
|
|
||
| #ifndef ELASTICNETPENALTY_H | ||
| #define ELASTICNETPENALTY_H | ||
| #include <shogun/optimization/SparsePenalty.h> | ||
| #include <shogun/optimization/L1Penalty.h> | ||
| #include <shogun/optimization/L2Penalty.h> | ||
| #include <shogun/lib/config.h> | ||
| #include <shogun/mathematics/Math.h> | ||
| namespace shogun | ||
| { | ||
| /** @brief The is the base class for ElasticNet penalty/regularization within the FirstOrderMinimizer framework. | ||
| * | ||
| * For ElasticNet penalty, \f$ElasticNet(w)\f$ | ||
| * \f[ | ||
| * ElasticNet(w)= \lambda \| w \|_1 + (1.0-\lambda) \| w \|_2 | ||
| * \f] | ||
| * where \f$\lambda\f$ is the l1_ratio. | ||
| */ | ||
|
|
||
| class ElasticNetPenalty: public SparsePenalty | ||
| { | ||
| public: | ||
| ElasticNetPenalty() | ||
| :SparsePenalty() {init();} | ||
|
|
||
| virtual ~ElasticNetPenalty() | ||
| { | ||
| delete m_l1_penalty; | ||
| delete m_l2_penalty; | ||
| } | ||
|
|
||
| virtual void set_l1_ratio(float64_t ratio) | ||
| { | ||
| REQUIRE(ratio>0.0 && ratio<1.0, ""); | ||
| m_l1_ratio=ratio; | ||
| } | ||
|
|
||
| /** Given the value of a target variable, | ||
| * this method returns the penalty of the variable | ||
| * | ||
| * @param variable value of the variable | ||
| * @return penalty of the variable | ||
| */ | ||
| virtual float64_t get_penalty(float64_t variable) | ||
| { | ||
| check_ratio(); | ||
| float64_t penalty=m_l1_ratio*m_l1_penalty->get_penalty(variable); | ||
| penalty+=(1.0-m_l1_ratio)*m_l2_penalty->get_penalty(variable); | ||
| return penalty; | ||
| } | ||
|
|
||
| virtual float64_t get_penalty_gradient(float64_t variable, | ||
| float64_t gradient_of_variable) | ||
| { | ||
| check_ratio(); | ||
| float64_t grad=m_l1_ratio*m_l1_penalty->get_penalty_gradient(variable, gradient_of_variable); | ||
| grad+=(1.0-m_l1_ratio)*m_l2_penalty->get_penalty_gradient(variable, gradient_of_variable); | ||
| return grad; | ||
| } | ||
|
|
||
| virtual void set_rounding_eplison(float64_t eplison) | ||
| { | ||
| m_l1_penalty->set_rounding_eplison(eplison); | ||
| } | ||
|
|
||
| virtual void update_sparse_variable(SGVector<float64_t> variable, | ||
| float64_t penalty_delta) | ||
| { | ||
| check_ratio(); | ||
| m_l1_penalty->update_sparse_variable(variable, penalty_delta*m_l1_ratio); | ||
| } | ||
|
|
||
| /** Update a context object to store mutable variables | ||
| * used in learning rate | ||
| * | ||
| * @param context, a context object | ||
| */ | ||
| virtual void update_context(CMinimizerContext* context) | ||
| { | ||
| REQUIRE(context, "Context must set\n"); | ||
| m_l1_penalty->update_context(context); | ||
| m_l2_penalty->update_context(context); | ||
| } | ||
|
|
||
| /** Load the given context object to restore mutable variables | ||
| * | ||
| * @param context, a context object | ||
| */ | ||
| virtual void load_from_context(CMinimizerContext* context) | ||
| { | ||
| REQUIRE(context, "Context must set\n"); | ||
| m_l1_penalty->load_from_context(context); | ||
| m_l2_penalty->load_from_context(context); | ||
| } | ||
| protected: | ||
| virtual void check_ratio() | ||
| { | ||
| REQUIRE(m_l1_ratio>0, "l1_ratio must set\n"); | ||
| } | ||
|
|
||
| float64_t m_l1_ratio; | ||
| L1Penalty* m_l1_penalty; | ||
| L2Penalty* m_l2_penalty; | ||
|
|
||
| private: | ||
| void init() | ||
| { | ||
| m_l1_ratio=0; | ||
| m_l1_penalty=new L1Penalty(); | ||
| m_l2_penalty=new L2Penalty(); | ||
| } | ||
| }; | ||
|
|
||
| } | ||
|
|
||
| #endif |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,129 @@ | ||
| /* | ||
| * Copyright (c) The Shogun Machine Learning Toolbox | ||
| * Written (w) 2015 Wu Lin | ||
| * All rights reserved. | ||
| * | ||
| * Redistribution and use in source and binary forms, with or without | ||
| * modification, are permitted provided that the following conditions are met: | ||
| * | ||
| * 1. Redistributions of source code must retain the above copyright notice, this | ||
| * list of conditions and the following disclaimer. | ||
| * 2. Redistributions in binary form must reproduce the above copyright notice, | ||
| * this list of conditions and the following disclaimer in the documentation | ||
| * and/or other materials provided with the distribution. | ||
| * | ||
| * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND | ||
| * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED | ||
| * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE | ||
| * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR | ||
| * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES | ||
| * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; | ||
| * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND | ||
| * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT | ||
| * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS | ||
| * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | ||
| * | ||
| * The views and conclusions contained in the software and documentation are those | ||
| * of the authors and should not be interpreted as representing official policies, | ||
| * either expressed or implied, of the Shogun Development Team. | ||
| * | ||
| */ | ||
|
|
||
| #ifndef L1PENALTY_H | ||
| #define L1PENALTY_H | ||
| #include <shogun/optimization/SparsePenalty.h> | ||
| #include <shogun/lib/config.h> | ||
| #include <shogun/mathematics/Math.h> | ||
| namespace shogun | ||
| { | ||
| /** @brief The is the base class for L1 penalty/regularization within the FirstOrderMinimizer framework. | ||
| * | ||
| * For L1 penalty, \f$L1(w)\f$ | ||
| * \f[ | ||
| * L1(w)=\| w \|_1 = \sum_i \| w_i \| | ||
| * \f] | ||
| */ | ||
|
|
||
| class L1Penalty: public SparsePenalty | ||
| { | ||
| public: | ||
| L1Penalty() | ||
| :SparsePenalty() {init();} | ||
|
|
||
| virtual ~L1Penalty() {} | ||
|
|
||
| /** Given the value of a target variable, | ||
| * this method returns the penalty of the variable | ||
| * | ||
| * @param variable value of the variable | ||
| * @return penalty of the variable | ||
| */ | ||
| virtual float64_t get_penalty(float64_t variable) {return CMath::abs(variable);} | ||
|
|
||
| virtual float64_t get_penalty_gradient(float64_t variable, | ||
| float64_t gradient_of_variable) {return 0.0;} | ||
|
|
||
| virtual void set_rounding_eplison(float64_t eplison) | ||
| { | ||
| REQUIRE(eplison>=0,"Rounding eplison (%f) should be non-negative\n", eplison); | ||
| m_rounding_eplison=eplison; | ||
| } | ||
|
|
||
| virtual void update_sparse_variable(SGVector<float64_t> variable, | ||
| float64_t penalty_delta) | ||
| { | ||
| for(index_t idx=0; idx<variable.vlen; idx++) | ||
| variable[idx]=get_sparse_variable(variable[idx], penalty_delta); | ||
| } | ||
|
|
||
| /** Update a context object to store mutable variables | ||
| * used in learning rate | ||
| * | ||
| * @param context, a context object | ||
| */ | ||
| virtual void update_context(CMinimizerContext* context) | ||
| { | ||
| REQUIRE(context, "Context must set\n"); | ||
| } | ||
|
|
||
| /** Load the given context object to restore mutable variables | ||
| * | ||
| * @param context, a context object | ||
| */ | ||
| virtual void load_from_context(CMinimizerContext* context) | ||
| { | ||
| REQUIRE(context, "Context must set\n"); | ||
| } | ||
| protected: | ||
| float64_t m_rounding_eplison; | ||
|
|
||
| virtual float64_t get_sparse_variable(float64_t variable, float64_t penalty_delta) | ||
| { | ||
| if (variable>0.0) | ||
| { | ||
| variable-=penalty_delta; | ||
| if (variable<0.0) | ||
| variable=0.0; | ||
| } | ||
| else | ||
| { | ||
| variable+=penalty_delta; | ||
| if (variable>0.0) | ||
| variable=0.0; | ||
| } | ||
| if (CMath::abs(variable)<m_rounding_eplison) | ||
| variable=0.0; | ||
| return variable; | ||
| return 0; | ||
| } | ||
|
|
||
| private: | ||
| void init() | ||
| { | ||
| m_rounding_eplison=1e-8; | ||
| } | ||
| }; | ||
|
|
||
| } | ||
|
|
||
| #endif |
Oops, something went wrong.