validate, nDCG

5 files changed, 109 insertions, 23 deletions

diff --git a/CMakeLists.txt b/CMakeLists.txt
index e1eb353..3d8a4e3 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -11,6 +11,6 @@ set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
 FIND_PACKAGE( Boost COMPONENTS program_options REQUIRED )
 INCLUDE_DIRECTORIES( ${Boost_INCLUDE_DIR})
 
-set(SOURCE_FILES main.cpp ./model/ranksvm.cpp ./model/ranksvmtn.cpp)
-add_executable(ranksvm ${SOURCE_FILES})
+set(SOURCE_FILES main.cpp ./model/ranksvm.cpp ./model/ranksvmtn.cpp ./model/rankaccu.cpp)
+add_executable(ranksvm ${SOURCE_FILES} model/rankaccu.h model/rankaccu.cpp)
 TARGET_LINK_LIBRARIES( ranksvm ${Boost_LIBRARIES} )
\ No newline at end of file
diff --git a/main.cpp b/main.cpp
index eeb6b99..0c71b07 100644
--- a/main.cpp
+++ b/main.cpp
@@ -5,8 +5,7 @@
 #include "tools/easylogging++.h"
 #include "model/ranksvmtn.h"
 #include "tools/fileDataProvider.h"
-#include "tools/matrixIO.h"
-#include <fstream>
+#include "model/rankaccu.h"
 
 INITIALIZE_EASYLOGGINGPP
 
@@ -28,6 +27,10 @@ int train() {
     dp.getDataSet(D);
     LOG(INFO)<<"Read "<<D.getSize()<<" entries with "<< D.getfSize()<<" features";
     rsvm->train(D);
+    std::vector<double> L;
+    rsvm->predict(D,L);
+
+    rank_accu(D,L);
 
     LOG(INFO)<<"Training finished,saving model";
 
@@ -54,24 +57,31 @@ int predict() {
         rsvm->predict(D,L);
     }
 
-    LOG(INFO)<<"Finished,saving prediction";
-    std::ofstream fout(vm["output"].as<std::string>().c_str());
+    if (vm.count("validate"))
+    {
+        rank_accu(D,L);
+    }
 
-    for (int i=0; i<L.size();++i)
-        fout<<L[i]<<std::endl;
-    fout.close();
+    if (vm.count("output"))
+    {
+        LOG(INFO)<<"Finished,saving prediction";
+        std::ofstream fout(vm["output"].as<std::string>().c_str());
 
+        for (int i=0; i<L.size();++i)
+            fout<<L[i]<<std::endl;
+        fout.close();
+    }
+    else if (!vm.count("validate"))
+    {
+        LOG(INFO)<<"Finished";
+        for (int i=0; i<L.size();++i)
+            std::cout<<L[i]<<std::endl;
+    }
     dp.close();
     delete rsvm;
     return 0;
 }
 
-int validate()
-{
-    LOG(FATAL)<<"Not Implemented";
-    return 0;
-}
-
 int main(int argc, char **argv) {
     // Defining program options
     po::options_description desc("Allowed options");
@@ -95,15 +105,9 @@ int main(int argc, char **argv) {
     }
 
     if (vm.count("train")) {
-        LOG(INFO) << "Program option: training";
         train();
     }
-    else if (vm.count("validate")) {
-        LOG(INFO) << "Program option: validate";
-        validate();
-    }
-    else if (vm.count("predict")) {
-        LOG(INFO) << "Program option: predict";
+    else if (vm.count("validate")||vm.count("predict")) {
         predict();
     }
     return 0;
diff --git a/model/rankaccu.cpp b/model/rankaccu.cpp
new file mode 100644
index 0000000..2e77eb6
--- /dev/null
+++ b/model/rankaccu.cpp
@@ -0,0 +1,69 @@
+//
+// Created by joe on 4/12/15.
+//
+
+#include "rankaccu.h"
+#include "../tools/easylogging++.h"
+
+using namespace std;
+
+const double offset = 1;
+
+void ranksort(int l,int r,vector<int> &rank,const vector<double> &ref1,const vector<double> &ref2)
+{
+    int i=l,j=r,k;
+    double mid1=ref1[rank[(l+r)>>1]],mid2=ref2[rank[(l+r)>>1]];
+    while (i<=j)
+    {
+        while (ref1[rank[i]]>mid1 || (ref1[rank[i]]==mid1 && ref2[rank[i]]>mid2)) ++i;
+        while (ref1[rank[j]]<mid1 || (ref1[rank[j]]==mid1 && ref2[rank[j]]<mid2)) --j;
+        if (i<=j)
+        {
+            k=rank[i];
+            rank[i]=rank[j];
+            rank[j]=k;
+            ++i;
+            --j;
+        }
+    }
+    if (j>l)
+        ranksort(l,j,rank,ref1,ref2);
+    if (i<r)
+        ranksort(i,r,rank,ref1,ref2);
+}
+
+int rank_accu(DataList &D,const vector<double> pred)
+{
+    unsigned long n = D.getSize();
+    vector<int> orig_rank(n),pred_rank(n);
+    vector<double> orig(n);
+    int i,j;
+    for (i=0;i<D.getSize();++i)
+    {
+        orig_rank[i]=i;
+        pred_rank[i]=i;
+        orig[i]=D.getData()[i]->rank;
+    }
+    int cnt=0;
+    double accu_nDCG=0;
+    i=j=0;
+    while (i<D.getSize())
+    {
+        if ((i+1 == D.getSize())|| D.getData()[i]->qid!=D.getData()[i+1]->qid)
+        {
+            double Y=0,Z=0;
+            ranksort(j,i,orig_rank,orig,pred);
+            ranksort(j,i,pred_rank,pred,orig);
+            for (int k = j;k<=i;++k)
+            {
+                Z += (pow(2,offset+orig[orig_rank[k]]) - 1)/log2(2+k-j);
+                Y += (pow(2,offset+orig[pred_rank[k]]) - 1)/log2(2+k-j);
+            }
+            accu_nDCG+=Y/Z;
+            j = i+1;
+            ++cnt;
+        }
+        ++i;
+    }
+    LOG(INFO)<<"Average nDGC over "<< cnt<< " queries: "<< accu_nDCG/cnt;
+}
\ No newline at end of file
diff --git a/model/rankaccu.h b/model/rankaccu.h
new file mode 100644
index 0000000..3fe5379
--- /dev/null
+++ b/model/rankaccu.h
@@ -0,0 +1,13 @@
+//
+// Created by joe on 4/12/15.
+//
+
+#ifndef RANKSVM_RANKACCU_H
+#define RANKSVM_RANKACCU_H
+
+#include<vector>
+#include"../tools/dataProvider.h"
+
+int rank_accu(DataList &D,const std::vector<double> pred);
+
+#endif //RANKSVM_RANKACCU_H
diff --git a/tools/dataProvider.h b/tools/dataProvider.h
index da3e1ee..2c3169a 100644
--- a/tools/dataProvider.h
+++ b/tools/dataProvider.h
@@ -35,7 +35,7 @@ public:
             delete data[i];
         data.clear();
     }
-    std::vector<DataEntry*>& getData(){
+    inline std::vector<DataEntry*>& getData(){
         return data;
     }
     ~DataList(){