Added a set of kernels that can operate on sparse vectors.

--HG--
extra : convert_revision : svn%3Afdd8eb12-d10e-0410-9acb-85c331704f74/trunk%402949

dlib/svm.h

@@ -13,6 +13,7 @@
 #include "svm/reduced.h"
 #include "svm/rvm.h"
 #include "svm/pegasos.h"
+#include "svm/sparse_kernel.h"
 
 #endif // DLIB_SVm_HEADER
 

dlib/svm/sparse_kernel.h

+// Copyright (C) 2009  Davis E. King (davisking@users.sourceforge.net)
+// License: Boost Software License   See LICENSE.txt for the full license.
+#ifndef DLIB_SVm_SPARSE_KERNEL
+#define DLIB_SVm_SPARSE_KERNEL
+
+#include "sparse_kernel_abstract.h"
+#include <cmath>
+#include <limits>
+#include "../algs.h"
+#include "../serialize.h"
+#include "sparse_vector.h"
+
+
+namespace dlib
+{
+
+// ----------------------------------------------------------------------------------------
+
+    template <
+        typename T
+        >
+    struct sparse_radial_basis_kernel
+    {
+        typedef typename T::value_type::second_type scalar_type;
+        typedef T sample_type;
+        typedef dlib::memory_manager<char>::kernel_1a mem_manager_type;
+
+        sparse_radial_basis_kernel(const scalar_type g) : gamma(g) {}
+        sparse_radial_basis_kernel() : gamma(0.1) {}
+        sparse_radial_basis_kernel(
+            const sparse_radial_basis_kernel& k
+        ) : gamma(k.gamma) {}
+
+
+        const scalar_type gamma;
+
+        scalar_type operator() (
+            const sample_type& a,
+            const sample_type& b
+        ) const
+        { 
+            const scalar_type d = sparse_vector::distance_squared(a,b);
+            return std::exp(-gamma*d);
+        }
+
+        sparse_radial_basis_kernel& operator= (
+            const sparse_radial_basis_kernel& k
+        )
+        {
+            const_cast<scalar_type&>(gamma) = k.gamma;
+            return *this;
+        }
+
+        bool operator== (
+            const sparse_radial_basis_kernel& k
+        ) const
+        {
+            return gamma == k.gamma;
+        }
+    };
+
+    template <
+        typename T
+        >
+    void serialize (
+        const sparse_radial_basis_kernel<T>& item,
+        std::ostream& out
+    )
+    {
+        try
+        {
+            serialize(item.gamma, out);
+        }
+        catch (serialization_error& e)
+        { 
+            throw serialization_error(e.info + "\n   while serializing object of type sparse_radial_basis_kernel"); 
+        }
+    }
+
+    template <
+        typename T
+        >
+    void deserialize (
+        sparse_radial_basis_kernel<T>& item,
+        std::istream& in 
+    )
+    {
+        typedef typename T::type scalar_type;
+        try
+        {
+            deserialize(const_cast<scalar_type&>(item.gamma), in);
+        }
+        catch (serialization_error& e)
+        { 
+            throw serialization_error(e.info + "\n   while deserializing object of type sparse_radial_basis_kernel"); 
+        }
+    }
+
+// ----------------------------------------------------------------------------------------
+
+    template <
+        typename T
+        >
+    struct sparse_polynomial_kernel
+    {
+        typedef typename T::value_type::second_type scalar_type;
+        typedef T sample_type;
+        typedef dlib::memory_manager<char>::kernel_1a mem_manager_type;
+
+        sparse_polynomial_kernel(const scalar_type g, const scalar_type c, const scalar_type d) : gamma(g), coef(c), degree(d) {}
+        sparse_polynomial_kernel() : gamma(1), coef(0), degree(1) {}
+        sparse_polynomial_kernel(
+            const sparse_polynomial_kernel& k
+        ) : gamma(k.gamma), coef(k.coef), degree(k.degree) {}
+
+        typedef T type;
+        const scalar_type gamma;
+        const scalar_type coef;
+        const scalar_type degree;
+
+        scalar_type operator() (
+            const sample_type& a,
+            const sample_type& b
+        ) const
+        { 
+            return std::pow(gamma*(sparse_vector::dot_product(a,b)) + coef, degree);
+        }
+
+        sparse_polynomial_kernel& operator= (
+            const sparse_polynomial_kernel& k
+        )
+        {
+            const_cast<scalar_type&>(gamma) = k.gamma;
+            const_cast<scalar_type&>(coef) = k.coef;
+            const_cast<scalar_type&>(degree) = k.degree;
+            return *this;
+        }
+
+        bool operator== (
+            const sparse_polynomial_kernel& k
+        ) const
+        {
+            return (gamma == k.gamma) && (coef == k.coef) && (degree == k.degree);
+        }
+    };
+
+    template <
+        typename T
+        >
+    void serialize (
+        const sparse_polynomial_kernel<T>& item,
+        std::ostream& out
+    )
+    {
+        try
+        {
+            serialize(item.gamma, out);
+            serialize(item.coef, out);
+            serialize(item.degree, out);
+        }
+        catch (serialization_error& e)
+        { 
+            throw serialization_error(e.info + "\n   while serializing object of type sparse_polynomial_kernel"); 
+        }
+    }
+
+    template <
+        typename T
+        >
+    void deserialize (
+        sparse_polynomial_kernel<T>& item,
+        std::istream& in 
+    )
+    {
+        typedef typename T::type scalar_type;
+        try
+        {
+            deserialize(const_cast<scalar_type&>(item.gamma), in);
+            deserialize(const_cast<scalar_type&>(item.coef), in);
+            deserialize(const_cast<scalar_type&>(item.degree), in);
+        }
+        catch (serialization_error& e)
+        { 
+            throw serialization_error(e.info + "\n   while deserializing object of type sparse_polynomial_kernel"); 
+        }
+    }
+
+// ----------------------------------------------------------------------------------------
+
+    template <
+        typename T
+        >
+    struct sparse_sigmoid_kernel
+    {
+        typedef typename T::value_type::second_type scalar_type;
+        typedef T sample_type;
+        typedef dlib::memory_manager<char>::kernel_1a mem_manager_type;
+
+        sparse_sigmoid_kernel(const scalar_type g, const scalar_type c) : gamma(g), coef(c) {}
+        sparse_sigmoid_kernel() : gamma(0.1), coef(-1.0) {}
+        sparse_sigmoid_kernel(
+            const sparse_sigmoid_kernel& k
+        ) : gamma(k.gamma), coef(k.coef) {}
+
+        typedef T type;
+        const scalar_type gamma;
+        const scalar_type coef;
+
+        scalar_type operator() (
+            const sample_type& a,
+            const sample_type& b
+        ) const
+        { 
+            return std::tanh(gamma*(sparse_vector::dot_product(a,b)) + coef);
+        }
+
+        sparse_sigmoid_kernel& operator= (
+            const sparse_sigmoid_kernel& k
+        )
+        {
+            const_cast<scalar_type&>(gamma) = k.gamma;
+            const_cast<scalar_type&>(coef) = k.coef;
+            return *this;
+        }
+
+        bool operator== (
+            const sparse_sigmoid_kernel& k
+        ) const
+        {
+            return (gamma == k.gamma) && (coef == k.coef);
+        }
+    };
+
+    template <
+        typename T
+        >
+    void serialize (
+        const sparse_sigmoid_kernel<T>& item,
+        std::ostream& out
+    )
+    {
+        try
+        {
+            serialize(item.gamma, out);
+            serialize(item.coef, out);
+        }
+        catch (serialization_error& e)
+        { 
+            throw serialization_error(e.info + "\n   while serializing object of type sparse_sigmoid_kernel"); 
+        }
+    }
+
+    template <
+        typename T
+        >
+    void deserialize (
+        sparse_sigmoid_kernel<T>& item,
+        std::istream& in 
+    )
+    {
+        typedef typename T::type scalar_type;
+        try
+        {
+            deserialize(const_cast<scalar_type&>(item.gamma), in);
+            deserialize(const_cast<scalar_type&>(item.coef), in);
+        }
+        catch (serialization_error& e)
+        { 
+            throw serialization_error(e.info + "\n   while deserializing object of type sparse_sigmoid_kernel"); 
+        }
+    }
+
+// ----------------------------------------------------------------------------------------
+
+    template <typename T>
+    struct sparse_linear_kernel
+    {
+        typedef typename T::value_type::second_type scalar_type;
+        typedef T sample_type;
+        typedef dlib::memory_manager<char>::kernel_1a mem_manager_type;
+
+        scalar_type operator() (
+            const sample_type& a,
+            const sample_type& b
+        ) const
+        { 
+            return sparse_vector::dot_product(a,b);
+        }
+
+        bool operator== (
+            const sparse_linear_kernel& k
+        ) const
+        {
+            return true;
+        }
+    };
+
+    template <
+        typename T
+        >
+    void serialize (
+        const sparse_linear_kernel<T>& item,
+        std::ostream& out
+    ){}
+
+    template <
+        typename T
+        >
+    void deserialize (
+        sparse_linear_kernel<T>& item,
+        std::istream& in 
+    ){}
+
+// ----------------------------------------------------------------------------------------
+
+}
+
+#endif // DLIB_SVm_SPARSE_KERNEL
+
+
+

dlib/svm/sparse_kernel_abstract.h

+// Copyright (C) 2009  Davis E. King (davisking@users.sourceforge.net)
+// License: Boost Software License   See LICENSE.txt for the full license.
+#undef DLIB_SVm_SPARSE_KERNEL_ABSTRACT_
+#ifdef DLIB_SVm_SPARSE_KERNEL_ABSTRACT_
+
+#include <cmath>
+#include <limits>
+#include "../algs.h"
+#include "../serialize.h"
+#include "kernel_abstract.h"
+#include "sparse_vector_abstract.h"
+
+namespace dlib
+{
+
+// ----------------------------------------------------------------------------------------
+
+    template <
+        typename T
+        >
+    struct sparse_radial_basis_kernel
+    {
+        /*!
+            REQUIREMENTS ON T
+                Must be a sparse vector as defined in dlib/svm/sparse_vector_abstract.h
+
+            WHAT THIS OBJECT REPRESENTS
+                This object represents a radial basis function kernel
+                that works with sparse vectors.
+        !*/
+
+        typedef typename T::value_type::second_type scalar_type;
+        typedef T sample_type;
+        typedef dlib::memory_manager<char>::kernel_1a mem_manager_type;
+
+        const scalar_type gamma;
+
+        sparse_radial_basis_kernel(
+        );
+        /*!
+            ensures
+                - #gamma == 0.1 
+        !*/
+
+        sparse_radial_basis_kernel(
+            const sparse_radial_basis_kernel& k
+        );
+        /*!
+            ensures
+                - #gamma == k.gamma
+        !*/
+
+        sparse_radial_basis_kernel(
+            const scalar_type g
+        );
+        /*!
+            ensures
+                - #gamma == g
+        !*/
+
+        scalar_type operator() (
+            const sample_type& a,
+            const sample_type& b
+        ) const;
+        /*!
+            requires
+                - a contains a sorted range 
+                - b contains a sorted range 
+            ensures
+                - returns exp(-gamma * sparse_vector::distance_squared(a,b))
+        !*/
+
+        sparse_radial_basis_kernel& operator= (
+            const sparse_radial_basis_kernel& k
+        );
+        /*!
+            ensures
+                - #gamma = k.gamma
+                - returns *this
+        !*/
+
+        bool operator== (
+            const sparse_radial_basis_kernel& k
+        ) const;
+        /*!
+            ensures
+                - if (k and *this are identical) then
+                    - returns true
+                - else
+                    - returns false
+        !*/
+
+    };
+
+    template <
+        typename T
+        >
+    void serialize (
+        const sparse_radial_basis_kernel<T>& item,
+        std::ostream& out
+    );
+    /*!
+        provides serialization support for sparse_radial_basis_kernel
+    !*/
+
+    template <
+        typename T
+        >
+    void deserialize (
+        sparse_radial_basis_kernel<T>& item,
+        std::istream& in 
+    );
+    /*!
+        provides deserialization support for sparse_radial_basis_kernel
+    !*/
+
+// ----------------------------------------------------------------------------------------
+
+    template <
+        typename T
+        >
+    struct sparse_sigmoid_kernel
+    {
+        /*!
+            REQUIREMENTS ON T
+                Must be a sparse vector as defined in dlib/svm/sparse_vector_abstract.h
+
+            WHAT THIS OBJECT REPRESENTS
+                This object represents a sigmoid kernel
+                that works with sparse vectors.
+        !*/
+
+        typedef typename T::value_type::second_type scalar_type;
+        typedef T sample_type;
+        typedef dlib::memory_manager<char>::kernel_1a mem_manager_type;
+
+        const scalar_type gamma;
+        const scalar_type coef;
+
+        sparse_sigmoid_kernel(
+        );
+        /*!
+            ensures
+                - #gamma == 0.1 
+                - #coef == -1.0 
+        !*/
+
+        sparse_sigmoid_kernel(
+            const sparse_sigmoid_kernel& k
+        );
+        /*!
+            ensures
+                - #gamma == k.gamma
+                - #coef == k.coef
+        !*/
+
+        sparse_sigmoid_kernel(
+            const scalar_type g,
+            const scalar_type c
+        );
+        /*!
+            ensures
+                - #gamma == g
+                - #coef == c
+        !*/
+
+        scalar_type operator() (
+            const sample_type& a,
+            const sample_type& b
+        ) const;
+        /*!
+            requires
+                - a contains a sorted range 
+                - b contains a sorted range 
+            ensures
+                - returns tanh(gamma * sparse_vector::dot(a,b) + coef)
+        !*/
+
+        sparse_sigmoid_kernel& operator= (
+            const sparse_sigmoid_kernel& k
+        );
+        /*!
+            ensures
+                - #gamma = k.gamma
+                - #coef = k.coef
+                - returns *this
+        !*/
+
+        bool operator== (
+            const sparse_sigmoid_kernel& k
+        ) const;
+        /*!
+            ensures
+                - if (k and *this are identical) then
+                    - returns true
+                - else
+                    - returns false
+        !*/
+    };
+
+    template <
+        typename T
+        >
+    void serialize (
+        const sparse_sigmoid_kernel<T>& item,
+        std::ostream& out
+    );
+    /*!
+        provides serialization support for sparse_sigmoid_kernel
+    !*/
+
+    template <
+        typename T
+        >
+    void deserialize (
+        sparse_sigmoid_kernel<T>& item,
+        std::istream& in 
+    );
+    /*!
+        provides deserialization support for sparse_sigmoid_kernel
+    !*/
+
+
+// ----------------------------------------------------------------------------------------
+
+    template <
+        typename T
+        >
+    struct sparse_polynomial_kernel
+    {
+        /*!
+            REQUIREMENTS ON T
+                Must be a sparse vector as defined in dlib/svm/sparse_vector_abstract.h
+
+            WHAT THIS OBJECT REPRESENTS
+                This object represents a polynomial kernel
+                that works with sparse vectors.
+        !*/
+
+        typedef typename T::value_type::second_type scalar_type;
+        typedef T sample_type;
+        typedef dlib::memory_manager<char>::kernel_1a mem_manager_type;
+
+        const scalar_type gamma;
+        const scalar_type coef;
+        const scalar_type degree;
+
+        sparse_polynomial_kernel(
+        );
+        /*!
+            ensures
+                - #gamma == 1 
+                - #coef == 0 
+                - #degree == 1 
+        !*/
+
+        sparse_polynomial_kernel(
+            const sparse_polynomial_kernel& k
+        );
+        /*!
+            ensures
+                - #gamma == k.gamma
+                - #coef == k.coef
+                - #degree == k.degree
+        !*/
+
+        sparse_polynomial_kernel(
+            const scalar_type g,
+            const scalar_type c,
+            const scalar_type d
+        );
+        /*!
+            ensures
+                - #gamma == g
+                - #coef == c
+                - #degree == d
+        !*/
+
+        scalar_type operator() (
+            const sample_type& a,
+            const sample_type& b
+        ) const;
+        /*!
+            requires
+                - a contains a sorted range 
+                - b contains a sorted range 
+            ensures
+                - returns pow(gamma * sparse_vector::dot(a,b) + coef, degree)
+        !*/
+
+        sparse_polynomial_kernel& operator= (
+            const sparse_polynomial_kernel& k
+        );
+        /*!
+            ensures
+                - #gamma = k.gamma
+                - #coef = k.coef
+                - #degree = k.degree
+                - returns *this
+        !*/
+
+        bool operator== (
+            const sparse_polynomial_kernel& k
+        ) const;
+        /*!
+            ensures
+                - if (k and *this are identical) then
+                    - returns true
+                - else
+                    - returns false
+        !*/
+    };
+
+    template <
+        typename T
+        >
+    void serialize (
+        const sparse_polynomial_kernel<T>& item,
+        std::ostream& out
+    );
+    /*!
+        provides serialization support for sparse_polynomial_kernel
+    !*/
+
+    template <
+        typename T
+        >
+    void deserialize (
+        sparse_polynomial_kernel<T>& item,
+        std::istream& in 
+    );
+    /*!
+        provides deserialization support for sparse_polynomial_kernel
+    !*/
+
+// ----------------------------------------------------------------------------------------
+
+    template <
+        typename T
+        >
+    struct sparse_linear_kernel
+    {
+        /*!
+            REQUIREMENTS ON T
+                Must be a sparse vector as defined in dlib/svm/sparse_vector_abstract.h
+
+            WHAT THIS OBJECT REPRESENTS
+                This object represents a linear function kernel
+                that works with sparse vectors.
+        !*/
+
+        typedef typename T::value_type::second_type scalar_type;
+        typedef T sample_type;
+        typedef dlib::memory_manager<char>::kernel_1a mem_manager_type;
+
+        scalar_type operator() (
+            const sample_type& a,
+            const sample_type& b
+        ) const;
+        /*!
+            requires
+                - a contains a sorted range 
+                - b contains a sorted range 
+            ensures
+                - returns sparse_vector::dot(a,b) 
+        !*/
+
+        bool operator== (
+            const sparse_linear_kernel& k
+        ) const;
+        /*!
+            ensures
+                - returns true
+        !*/
+    };
+
+    template <
+        typename T
+        >
+    void serialize (
+        const sparse_linear_kernel<T>& item,
+        std::ostream& out
+    );
+    /*!
+        provides serialization support for sparse_linear_kernel
+    !*/
+
+    template <
+        typename T
+        >
+    void deserialize (
+        sparse_linear_kernel<T>& item,
+        std::istream& in 
+    );
+    /*!
+        provides deserialization support for sparse_linear_kernel 
+    !*/
+
+// ----------------------------------------------------------------------------------------
+
+}
+
+#endif // DLIB_SVm_SPARSE_KERNEL_ABSTRACT_
+
+

dlib/svm/sparse_vector.h

+// Copyright (C) 2009  Davis E. King (davisking@users.sourceforge.net)
+// License: Boost Software License   See LICENSE.txt for the full license.
+#ifndef DLIB_SVm_SPARSE_VECTOR
+#define DLIB_SVm_SPARSE_VECTOR
+
+#include "sparse_vector_abstract.h"
+#include <cmath>
+#include <limits>
+#include "../algs.h"
+
+
+namespace dlib
+{
+
+// ----------------------------------------------------------------------------------------
+
+    namespace sparse_vector
+    {
+
+        template <typename T, typename U>
+        typename T::value_type::second_type distance_squared (
+            const T& a,
+            const U& b
+        )
+        {
+            typedef typename T::value_type::second_type scalar_type;
+            typedef typename U::value_type::second_type scalar_typeU;
+            // Both T and U must contain the same kinds of elements
+            COMPILE_TIME_ASSERT((is_same_type<scalar_type, scalar_typeU>::value));
+
+            typename T::const_iterator ai = a.begin();
+            typename U::const_iterator bi = b.begin();
+
+            scalar_type sum = 0, temp = 0;
+            while (ai != a.end() && bi != b.end())
+            {
+                if (ai->first == bi->first)
+                {
+                    temp = ai->second - bi->second;
+                    ++ai;
+                    ++bi;
+                }
+                else if (ai->first < bi->first)
+                {
+                    temp = ai->second;
+                    ++ai;
+                }
+                else 
+                {
+                    temp = bi->second;
+                    ++bi;
+                }
+
+                sum += temp*temp;
+            }
+
+            while (ai != a.end())
+            {
+                sum += ai->second*ai->second;
+            }
+            while (bi != b.end())
+            {
+                sum += bi->second*bi->second;
+            }
+
+            return sum;
+        }
+
+    // ------------------------------------------------------------------------------------
+
+        template <typename T, typename U>
+        typename T::value_type::second_type dot_product (
+            const T& a,
+            const U& b
+        )
+        {
+            typedef typename T::value_type::second_type scalar_type;
+            typedef typename U::value_type::second_type scalar_typeU;
+            // Both T and U must contain the same kinds of elements
+            COMPILE_TIME_ASSERT((is_same_type<scalar_type, scalar_typeU>::value));
+
+            typename T::const_iterator ai = a.begin();
+            typename U::const_iterator bi = b.begin();
+
+            scalar_type sum = 0;
+            while (ai != a.end() && bi != b.end())
+            {
+                if (ai->first == bi->first)
+                {
+                    sum += ai->second * bi->second;
+                    ++ai;
+                    ++bi;
+                }
+                else if (ai->first < bi->first)
+                {
+                    ++ai;
+                }
+                else 
+                {
+                    ++bi;
+                }
+            }
+
+            return sum;
+        }
+
+    // ------------------------------------------------------------------------------------
+
+        template <typename T>
+        typename T::value_type::second_type length_squared (
+            const T& a
+        )
+        {
+            typedef typename T::value_type::second_type scalar_type;
+
+            typename T::const_iterator i = a.begin();
+
+            scalar_type sum = 0;
+            while (i != a.end())
+            {
+                sum += i->second * i->second;
+            }
+
+            return sum;
+        }
+
+    }
+
+// ----------------------------------------------------------------------------------------
+
+}
+
+#endif // DLIB_SVm_SPARSE_VECTOR
+
+
+
+

dlib/svm/sparse_vector_abstract.h

+// Copyright (C) 2009  Davis E. King (davisking@users.sourceforge.net)
+// License: Boost Software License   See LICENSE.txt for the full license.
+#undef DLIB_SVm_SPARSE_VECTOR_ABSTRACT_
+#ifdef DLIB_SVm_SPARSE_VECTOR_ABSTRACT_
+
+#include <cmath>
+#include "../algs.h"
+#include "../serialize.h"
+
+namespace dlib
+{
+
+// ----------------------------------------------------------------------------------------
+
+    /*!
+        In dlib, sparse vectors are represented using the container objects
+        in the C++ STL.  In particular, a sparse vector is any container that 
+        contains a sorted range of std::pair<key, scalar_value> objects where:
+            - key is any type that can serve as a unique index or identifier (e.g. long)
+            - scalar_value is float, double, or long double
+
+        So examples of valid sparse vectors are:    
+            - std::map<long, double>
+            - std::vector<std::pair<long, float> > where the vector is sorted.
+              (you could make sure it was sorted by applying std::sort to it)
+
+
+        This file defines a number of helper functions for doing normal vector
+        arithmetic things with sparse vectors.
+    !*/
+
+// ----------------------------------------------------------------------------------------
+
+    namespace sparse_vector
+    {
+        template <typename T, typename U>
+        typename T::value_type::second_type distance_squared (
+            const T& a,
+            const U& b
+        );
+        /*!
+            requires
+                - a is a sorted range of std::pair objects
+                - b is a sorted range of std::pair objects
+            ensures
+                - returns the squared distance between the vectors
+                  a and b
+        !*/
+
+    // ----------------------------------------------------------------------------------------
+
+        template <typename T, typename U>
+        typename T::value_type::second_type dot_product (
+            const T& a,
+            const U& b
+        );
+        /*!
+            requires
+                - a is a sorted range of std::pair objects
+                - b is a sorted range of std::pair objects
+            ensures
+                - returns the dot product between the vectors a and b
+        !*/
+
+    // ----------------------------------------------------------------------------------------
+
+        template <typename T>
+        typename T::value_type::second_type length_squared (
+            const T& a
+        );
+        /*!
+            requires
+                - a is a sorted range of std::pair objects
+                - b is a sorted range of std::pair objects
+            ensures
+                - returns dot(a,a) 
+        !*/
+
+    }
+
+// ----------------------------------------------------------------------------------------
+
+}
+
+#endif // DLIB_SVm_SPARSE_VECTOR_ABSTRACT_
+
+
+