From 82eadc91f0f697103e965bd3600475df60e8282f Mon Sep 17 00:00:00 2001
From: Peter Wu <peter@lekensteyn.nl>
Date: Wed, 1 Apr 2015 02:54:48 +0200
Subject: Calculate noisy channel probability

---
 spellchecker/src/ConfusionMatrixReader.java | 11 +++++++++++
 spellchecker/src/SpellCorrector.java        | 18 ++++++++++++++++++
 2 files changed, 29 insertions(+)

diff --git a/spellchecker/src/ConfusionMatrixReader.java b/spellchecker/src/ConfusionMatrixReader.java
index c9e79ab..90a8cef 100644
--- a/spellchecker/src/ConfusionMatrixReader.java
+++ b/spellchecker/src/ConfusionMatrixReader.java
@@ -60,4 +60,15 @@ public class ConfusionMatrixReader {
         Integer count = confusionMatrix.get(error + "|" + correct);
         return count == null ? 0 : count;
     }
+
+    /**
+     * Given the error {@code error}, find the number of occurrences of the
+     * error.
+     *
+     * @param error
+     * @return
+     */
+    public int getErrorsCount(String error) {
+        return countMatrix.getOrDefault(error, 0);
+    }
 }
diff --git a/spellchecker/src/SpellCorrector.java b/spellchecker/src/SpellCorrector.java
index 6c9225b..3fbd8cc 100644
--- a/spellchecker/src/SpellCorrector.java
+++ b/spellchecker/src/SpellCorrector.java
@@ -53,7 +53,25 @@ public class SpellCorrector {
                 return;
             }
 
+            // Find the channel model probability (probability of the edit).
+            // P(x|w) = "corrections count given error" / "errors count"
+            double correctionCount, errorCount, p_channel;
+            correctionCount = (double) cmr.getConfusionCount(error, correct);
+            errorCount = cmr.getErrorsCount(error);
+
+            // is this a known correction?
+            if (errorCount == 0.0) {
+                // no,
+                p_channel = 0.0;
+            } else {
+                // yes,
+                p_channel = correctionCount / errorCount;
+            }
+
+            // Sum the probabilities as independent modifications can result in
+            // the same word ("acess" -> "access" by "a|ac", "e|ce").
             double p = candidates.getOrDefault(word2, 0.0);
+            p += p_channel;
 
             candidates.put(word2, p);
         };
-- 
cgit v1.2.1