Move all prediction transform to the objective. (#383)

* many refactors.

* remove multi_loglossova.

* fix tests.

* avoid using lambda function.

* fix some format.

* reduce branching.

src/objective/objective_function.cpp

@@ -29,4 +29,30 @@ ObjectiveFunction* ObjectiveFunction::CreateObjectiveFunction(const std::string&
   }
   return nullptr;
 }
+
+ObjectiveFunction* ObjectiveFunction::CreateObjectiveFunction(const std::string& str) {
+  auto strs = Common::Split(str.c_str(), " ");
+  auto type = strs[0];
+  if (type == std::string("regression")) {
+    return new RegressionL2loss(strs);
+  } else if (type == std::string("regression_l1")) {
+    return new RegressionL1loss(strs);
+  } else if (type == std::string("huber")) {
+    return new RegressionHuberLoss(strs);
+  } else if (type == std::string("fair")) {
+    return new RegressionFairLoss(strs);
+  } else if (type == std::string("poisson")) {
+    return new RegressionPoissonLoss(strs);
+  } else if (type == std::string("binary")) {
+    return new BinaryLogloss(strs);
+  } else if (type == std::string("lambdarank")) {
+    return new LambdarankNDCG(strs);
+  } else if (type == std::string("multiclass")) {
+    return new MulticlassSoftmax(strs);
+  } else if (type == std::string("multiclassova")) {
+    return new MulticlassOVA(strs);
+  }
+  return nullptr;
+}
+
 }  // namespace LightGBM

src/objective/rank_objective.hpp

@@ -35,6 +35,11 @@ public:
       Log::Fatal("Sigmoid param %f should be greater than zero", sigmoid_);
     }
   }
+
+  explicit LambdarankNDCG(const std::vector<std::string>&) {
+
+  }
+
   ~LambdarankNDCG() {
 
   }
@@ -196,6 +201,12 @@ public:
     return "lambdarank";
   }
 
+  std::string ToString() const override {
+    std::stringstream str_buf;
+    str_buf << GetName();
+    return str_buf.str();
+  }
+
 private:
   /*! \brief Gains for labels */
   std::vector<double> label_gain_;

src/objective/regression_objective.hpp

@@ -13,6 +13,10 @@ public:
   explicit RegressionL2loss(const ObjectiveConfig&) {
   }
 
+  explicit RegressionL2loss(const std::vector<std::string>&) {
+
+  }
+
   ~RegressionL2loss() {
   }
 
@@ -43,6 +47,12 @@ public:
     return "regression";
   }
 
+  std::string ToString() const override {
+    std::stringstream str_buf;
+    str_buf << GetName();
+    return str_buf.str();
+  }
+
   bool IsConstantHessian() const override {
     if (weights_ == nullptr) {
       return true;
@@ -51,6 +61,8 @@ public:
     }
   }
 
+  bool BoostFromAverage() const override { return true; }
+
 private:
   /*! \brief Number of data */
   data_size_t num_data_;
@@ -69,6 +81,10 @@ public:
     eta_ = static_cast<double>(config.gaussian_eta);
   }
 
+  explicit RegressionL1loss(const std::vector<std::string>&) {
+
+  }
+
   ~RegressionL1loss() {}
 
   void Init(const Metadata& metadata, data_size_t num_data) override {
@@ -108,6 +124,14 @@ public:
     return "regression_l1";
   }
 
+  std::string ToString() const override {
+    std::stringstream str_buf;
+    str_buf << GetName();
+    return str_buf.str();
+  }
+
+  bool BoostFromAverage() const override { return true; }
+
 private:
   /*! \brief Number of data */
   data_size_t num_data_;
@@ -129,6 +153,10 @@ public:
     eta_ = static_cast<double>(config.gaussian_eta);
   }
 
+  explicit RegressionHuberLoss(const std::vector<std::string>&) {
+
+  }
+
   ~RegressionHuberLoss() {
   }
 
@@ -181,6 +209,14 @@ public:
     return "huber";
   }
 
+  std::string ToString() const override {
+    std::stringstream str_buf;
+    str_buf << GetName();
+    return str_buf.str();
+  }
+
+  bool BoostFromAverage() const override { return true; }
+
 private:
   /*! \brief Number of data */
   data_size_t num_data_;
@@ -202,6 +238,10 @@ public:
     c_ = static_cast<double>(config.fair_c);
   }
 
+  explicit RegressionFairLoss(const std::vector<std::string>&) {
+
+  }
+
   ~RegressionFairLoss() {}
 
   void Init(const Metadata& metadata, data_size_t num_data) override {
@@ -233,6 +273,14 @@ public:
     return "fair";
   }
 
+  std::string ToString() const override {
+    std::stringstream str_buf;
+    str_buf << GetName();
+    return str_buf.str();
+  }
+
+  bool BoostFromAverage() const override { return true; }
+
 private:
   /*! \brief Number of data */
   data_size_t num_data_;
@@ -254,6 +302,10 @@ public:
     max_delta_step_ = static_cast<double>(config.poisson_max_delta_step);
   }
 
+  explicit RegressionPoissonLoss(const std::vector<std::string>&) {
+
+  }
+
   ~RegressionPoissonLoss() {}
 
   void Init(const Metadata& metadata, data_size_t num_data) override {
@@ -283,6 +335,14 @@ public:
     return "poisson";
   }
 
+  std::string ToString() const override {
+    std::stringstream str_buf;
+    str_buf << GetName();
+    return str_buf.str();
+  }
+
+  bool BoostFromAverage() const override { return true; }
+
 private:
   /*! \brief Number of data */
   data_size_t num_data_;

windows/LightGBM.vcxproj

@@ -222,6 +222,7 @@
     <ClInclude Include="..\src\io\parser.hpp" />
     <ClInclude Include="..\src\io\sparse_bin.hpp" />
     <ClInclude Include="..\src\metric\binary_metric.hpp" />
+    <ClInclude Include="..\src\metric\map_metric.hpp" />
     <ClInclude Include="..\src\metric\rank_metric.hpp" />
     <ClInclude Include="..\src\metric\regression_metric.hpp" />
     <ClInclude Include="..\src\metric\multiclass_metric.hpp" />

windows/LightGBM.vcxproj.filters

@@ -180,6 +180,9 @@
     <ClInclude Include="..\include\LightGBM\utils\openmp_wrapper.h">
       <Filter>include\LightGBM\utils</Filter>
     </ClInclude>
+    <ClInclude Include="..\src\metric\map_metric.hpp">
+      <Filter>src\metric</Filter>
+    </ClInclude>
   </ItemGroup>
   <ItemGroup>
     <ClCompile Include="..\src\application\application.cpp">