--cumul now computes P(X <= x) instead of P(X < x).

[data-tool.git] / data-tool.cc

diff --git a/data-tool.cc b/data-tool.cc
index bcf19b3..9ab7a8c 100644 (file)
--- a/data-tool.cc
+++ b/data-tool.cc
@@ -92,6 +92,7 @@ void print_help_and_exit(int e) {
        << "  --help" << endl
        << "  --roc" << endl
        << "  --roc-surface" << endl
+       << "  --error" << endl
        << "  --normalize" << endl
        << "  --histo" << endl
        << "  --cumul" << endl
@@ -121,7 +122,7 @@ int main(int argc, char **argv) {
 
   int i = 1;
 
-  enum { UNKNOWN, ROC, ROC_SURFACE, HISTO, CUMUL, MISC } processing = UNKNOWN;
+  enum { UNKNOWN, ROC, ROC_SURFACE, ERROR, HISTO, CUMUL, MISC } processing = UNKNOWN;
 
   // Parsing the command line arguments ////////////////////////////////
 
@@ -141,6 +142,12 @@ int main(int argc, char **argv) {
       i++;
     }
 
+    else if(strcmp(argv[i], "--error") == 0) {
+      check_single_processing(processing == UNKNOWN);
+      processing = ERROR;
+      i++;
+    }
+
     else if(strcmp(argv[i], "--cumul") == 0) {
       check_single_processing(processing == UNKNOWN);
       processing = CUMUL;
@@ -233,7 +240,7 @@ int main(int argc, char **argv) {
         }
       }
 
-      Couple tmp[nb_samples];
+      Couple *tmp = new Couple[nb_samples];
       for(int n = 0; n < nb_samples; n++) {
         tmp[n].index = n;
         tmp[n].value = x[n];
@@ -242,8 +249,9 @@ int main(int argc, char **argv) {
       qsort(tmp, nb_samples, sizeof(Couple), compare_couple);
 
       for(int n = 0; n < nb_samples; n++)
-        cout << tmp[n].value << " " << double(n)/double(nb_samples)  << endl;
+        cout << tmp[n].value << " " << double(n+1)/double(nb_samples)  << endl;
 
+      delete[] tmp;
       delete[] x;
 
     }
@@ -252,6 +260,7 @@ int main(int argc, char **argv) {
 
   case ROC:
   case ROC_SURFACE:
+  case ERROR:
 
     {
       int nb_samples = 0, nb_samples_max = 1000;
@@ -276,7 +285,7 @@ int main(int argc, char **argv) {
         }
       }
 
-      Couple tmp[nb_samples];
+      Couple *tmp = new Couple[nb_samples];
       int nb_rn = 0, nb_rp = 0, nb_fp = 0, nb_fn = 0;
 
       bool binary = true;
@@ -305,11 +314,11 @@ int main(int argc, char **argv) {
           if(tmp[n].value < tmp[n+1].value) {
             cout << double(nb_fp)/double(nb_rn) << " "
                  << 1 - double(nb_fn) / double(nb_rp) << " "
-                 << (tmp[n].value + tmp[n+1].value) << " "
+                 << (tmp[n].value + tmp[n+1].value)/2 << " "
                  << endl;
           }
         }
-      } else {
+      } else if(processing == ROC_SURFACE) {
         double surface = 0;
         double cx = double(nb_fp)/double(nb_rn), cy = 1 - double(nb_fn) / double(nb_rp);
         for(int n = 0; n < nb_samples - 1; n++) {
@@ -322,9 +331,21 @@ int main(int argc, char **argv) {
           }
         }
         cout << surface  << endl;
+      } else {
+        for(int n = 0; n < nb_samples - 1; n++) {
+          if(x[tmp[n].index] >= 0) nb_fn++;
+          else                     nb_fp--;
+          if(tmp[n].value < tmp[n+1].value) {
+            cout << (tmp[n].value + tmp[n+1].value)/2 << " "
+                 << double(nb_fp + nb_fn)/double(nb_rn + nb_rp) << " "
+                 << endl;
+          }
+        }
       }
 
-      delete[] x; delete[] y;
+      delete[] tmp;
+      delete[] x;
+      delete[] y;
 
     }
 
@@ -356,7 +377,7 @@ int main(int argc, char **argv) {
         }
       }
 
-      int nb[nb_bins];
+      int *nb = new int[nb_bins];
       for(int n = 0; n < nb_bins; n++) nb[n] = 0;
 
       int nb_total = 0;
@@ -378,6 +399,8 @@ int main(int argc, char **argv) {
           cout << xmin + ((xmax - xmin) * n) / double(nb_bins) << " "
                << nb[n] / double(nb_total)  << endl;
       }
+
+      delete[] nb;
     }
 
     break;
@@ -412,7 +435,7 @@ int main(int argc, char **argv) {
         }
       }
 
-      Couple tmp[nb_samples];
+      Couple *tmp = new Couple[nb_samples];
       for(int n = 0; n < nb_samples; n++) {
         tmp[n].index = n;
         tmp[n].value = x[n];
@@ -437,6 +460,7 @@ int main(int argc, char **argv) {
            << " QUANTILE0.9 " << tmp[int(nb_samples * 0.9)].value
            << endl;
 
+      delete[] tmp;
     }
 
     break;