Changeset 500

Timestamp:

Jan 28, 2006, 5:41:54 PM (17 years ago)

Author:

Peter

Message:

implemented median function for gslapi::vector

changed sister function for std::vector to be named median_TMP because
the interface should be changed

Location:

trunk

Files:

• lib/classifier/ConsensusInputRanker.cc (modified) (1 diff)
• lib/statistics/WilcoxonFoldChange.cc (modified) (1 diff)
• lib/statistics/utility.cc (modified) (2 diffs)
• lib/statistics/utility.h (modified) (3 diffs)
• test/statistics_test.cc (modified) (1 diff)

trunk/lib/classifier/ConsensusInputRanker.cc

@@ -43 +43 @@
         ranks[j]=input_rankers_[j].rank(i);
       }
-      medians[i].first = statistics::median(ranks);
+      medians[i].first = statistics::median_TMP(ranks);
       medians[i].second = i;
     }

trunk/lib/statistics/WilcoxonFoldChange.cc

@@ -35 +35 @@
     }
     if (absolute_)
-      return fabs(median(distance));
-    return median(distance);
+      return fabs(median_TMP(distance));
+    return median_TMP(distance);
   }
 

trunk/lib/statistics/utility.cc

@@ -5 +5 @@
 
 #include <gsl/gsl_randist.h>
-
+#include <gsl/gsl_statistics_double.h>
 
 namespace theplu {
@@ -18 +18 @@
   }
 
+  double median(const gslapi::vector& vec, const bool sorted)
+  {
+    if (!sorted){
+      gslapi::vector vec_copy(vec);
+      vec_copy.sort();
+      return gsl_stats_median_from_sorted_data (vec_copy.gsl_vector_p()->data, 
+                                                vec_copy.gsl_vector_p()->stride,
+                                                vec_copy.gsl_vector_p()->size);
+    }
+    return gsl_stats_median_from_sorted_data (vec.gsl_vector_p()->data, 
+                                              vec.gsl_vector_p()->stride,
+                                              vec.gsl_vector_p()->size);
+  }
+
 }} // of namespace statistics and namespace theplu

trunk/lib/statistics/utility.h

@@ -3 +3 @@
 #ifndef _theplu_statistics_utility_ 
 #define _theplu_statistics_utility_ 
+
+#include <c++_tools/gslapi/vector.h>
 
 #include <algorithm>
@@ -35 +37 @@
   /// @return \a p'th percentile
   ///
+  /// @note interface will change
+  ///
   template <class T>
-  double percentile(std::vector<T>& vec, double p)
+  double percentile_TMP(std::vector<T>& vec, double p)
   {
     if (p==100)
@@ -52 +56 @@
   /// values. @return median
   ///
+  /// @note interface will change
+  ///
   template <class T> 
-  inline double median(std::vector<T>& v) { return percentile(v,50.0); }
+  inline double median_TMP(std::vector<T>& v) { return percentile_TMP(v,50.0); }
 
-  
+  ///
+  /// Median is defined to be value in the middle. If number of values
+  /// is even median is the average of the two middle values. If @a
+  /// sorted is true, the function assumes vector @a vec to be
+  /// sorted. If @a sorted is false, the vector is copied, the copy
+  /// is sorted, and then used to calculate the median.
+  ///
+  /// @return median
+  ///
+  double median(const gslapi::vector& vec, const bool sorted=true);
+
 }} // of namespace statistics and namespace theplu
 

trunk/test/statistics_test.cc

@@ -15 +15 @@
   for (unsigned int i=0; i<10; i++)
     data.push_back(static_cast<double>(i));
-  double m=theplu::statistics::median(data);
+  double m=theplu::statistics::median_TMP(data);
   if (m!=4.5)
     return -1;