goes from C-style loop variable to C++-style for trainAuto

modules/ml/src/svm.cpp

@@ -1671,19 +1671,17 @@ class SVMImpl : public SVM
             Mat temp_train_responses(train_sample_count, 1, rtype);
             Mat temp_test_responses;
 
-            int i,j,k,p;
-
-            for( p = range.start; p < range.end; p++ )
+            for( int p = range.start; p < range.end; p++ )
             {
                 svm->setParams(parameters[p]);
 
                 double error = 0;
-                for( k = 0; k < k_fold; k++ )
+                for( int k = 0; k < k_fold; k++ )
                 {
                     int start = (k*sample_count + k_fold/2)/k_fold;
-                    for( i = 0; i < train_sample_count; i++ )
+                    for( int i = 0; i < train_sample_count; i++ )
                     {
-                        j = sidx[(i+start)%sample_count];
+                        int j = sidx[(i+start)%sample_count];
                         memcpy(temp_train_samples.ptr(i), samples.ptr(j), sample_size);
                         if( is_classification )
                             temp_train_responses.at<int>(i) = responses.at<int>(j);
@@ -1695,17 +1693,17 @@ class SVMImpl : public SVM
                     if( !svm->do_train( temp_train_samples, temp_train_responses ))
                         continue;
 
-                    for( i = 0; i < test_sample_count; i++ )
+                    for( int i = 0; i < test_sample_count; i++ )
                     {
-                        j = sidx[(i+start+train_sample_count) % sample_count];
+                        int j = sidx[(i+start+train_sample_count) % sample_count];
                         memcpy(temp_test_samples.ptr(i), samples.ptr(j), sample_size);
                     }
 
                     svm->predict(temp_test_samples, temp_test_responses, 0);
-                    for( i = 0; i < test_sample_count; i++ )
+                    for( int i = 0; i < test_sample_count; i++ )
                     {
                         float val = temp_test_responses.at<float>(i);
-                        j = sidx[(i+start+train_sample_count) % sample_count];
+                        int j = sidx[(i+start+train_sample_count) % sample_count];
                         if( is_classification )
                             error += (float)(val != responses.at<int>(j));
                         else
@@ -1812,10 +1810,8 @@ class SVMImpl : public SVM
         vector<int> sidx;
         setRangeVector(sidx, sample_count);
 
-        int i, k;
-
         // randomly permute training samples
-        for( i = 0; i < sample_count; i++ )
+        for( int i = 0; i < sample_count; i++ )
         {
             int i1 = rng.uniform(0, sample_count);
             int i2 = rng.uniform(0, sample_count);
@@ -1829,7 +1825,7 @@ class SVMImpl : public SVM
             // between the k_fold parts.
             vector<int> sidx0, sidx1;
 
-            for( i = 0; i < sample_count; i++ )
+            for( int i = 0; i < sample_count; i++ )
             {
                 if( responses.at<int>(sidx[i]) == 0 )
                     sidx0.push_back(sidx[i]);
@@ -1840,15 +1836,15 @@ class SVMImpl : public SVM
             int n0 = (int)sidx0.size(), n1 = (int)sidx1.size();
             int a0 = 0, a1 = 0;
             sidx.clear();
-            for( k = 0; k < k_fold; k++ )
+            for( int k = 0; k < k_fold; k++ )
             {
                 int b0 = ((k+1)*n0 + k_fold/2)/k_fold, b1 = ((k+1)*n1 + k_fold/2)/k_fold;
                 int a = (int)sidx.size(), b = a + (b0 - a0) + (b1 - a1);
-                for( i = a0; i < b0; i++ )
+                for( int i = a0; i < b0; i++ )
                     sidx.push_back(sidx0[i]);
-                for( i = a1; i < b1; i++ )
+                for( int i = a1; i < b1; i++ )
                     sidx.push_back(sidx1[i]);
-                for( i = 0; i < (b - a); i++ )
+                for( int i = 0; i < (b - a); i++ )
                 {
                     int i1 = rng.uniform(a, b);
                     int i2 = rng.uniform(a, b);
@@ -1882,7 +1878,7 @@ class SVMImpl : public SVM
         // Extract the best parameters
         SvmParams best_params = params;
         double min_error = FLT_MAX;
-        for( i = 0; i < (int)result.size(); i++ )
+        for( int i = 0; i < (int)result.size(); i++ )
         {
             if( result[i] < min_error )
             {