From 0ee6610cf1c0ab342e438879b8a5fe4046e33189 Mon Sep 17 00:00:00 2001
From: Victor Chahuneau <vchahune@cs.cmu.edu>
Date: Wed, 25 Jul 2012 23:57:05 -0400
Subject: Character error rate metric

- [python] cdec.score.CER
- Cleanup .gitignore
---
 mteval/ns_cer.cc | 55 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 55 insertions(+)
 create mode 100644 mteval/ns_cer.cc

(limited to 'mteval/ns_cer.cc')

diff --git a/mteval/ns_cer.cc b/mteval/ns_cer.cc
new file mode 100644
index 00000000..a843d471
--- /dev/null
+++ b/mteval/ns_cer.cc
@@ -0,0 +1,55 @@
+#include "ns_cer.h"
+#include "tdict.h"
+
+static const unsigned kNUMFIELDS = 2;
+static const unsigned kEDITDISTANCE = 0;
+static const unsigned kCHARCOUNT = 1;
+
+bool CERMetric::IsErrorMetric() const {
+  return true;
+}
+
+unsigned CERMetric::SufficientStatisticsVectorSize() const {
+  return 2;
+}
+
+unsigned CERMetric::EditDistance(const std::string& hyp,
+                                 const std::string& ref) const {
+  const unsigned m = hyp.size(), n = ref.size();
+  std::vector<unsigned> edit((m + 1) * 2);
+  for(unsigned i = 0; i < n + 1; i++) {
+    for(unsigned j = 0; j < m + 1; j++) {
+      if(i == 0)
+        edit[j] = j;
+      else if(j == 0)
+        edit[(i%2)*(m+1)] = i;
+      else
+        edit[(i%2)*(m+1) + j] = std::min(std::min(edit[(i%2)*(m+1) + j-1] + 1,
+                                                   edit[((i-1)%2)*(m+1) + j] + 1),
+                                                   edit[((i-1)%2)*(m+1) + (j-1)] 
+                                                   + (hyp[j-1] == ref[i-1] ? 0 : 1));
+      
+    }
+  }
+  return edit[(n%2)*(m+1) + m];
+}
+
+void CERMetric::ComputeSufficientStatistics(const std::vector<WordID>& hyp,
+                                            const std::vector<std::vector<WordID> >& refs,
+                                            SufficientStats* out) const {
+  out->fields.resize(kNUMFIELDS);
+  std::string hyp_str(TD::GetString(hyp));
+  float best_score = hyp_str.size();
+  for (size_t i = 0; i < refs.size(); ++i) {
+    std::string ref_str(TD::GetString(refs[i]));
+    float score = EditDistance(hyp_str, ref_str);
+    if (score < best_score) {
+      out->fields[kEDITDISTANCE] = score;
+      out->fields[kCHARCOUNT] = ref_str.size();
+      best_score = score;
+    }
+  }
+}
+float CERMetric::ComputeScore(const SufficientStats& stats) const {
+  return stats.fields[kEDITDISTANCE] / stats.fields[kCHARCOUNT];
+}
-- 
cgit v1.2.3