diff options
| author |   satok <satok@google.com> | 2011-04-12 20:22:10 -0700 |
|---|---|---|
| committer |   Android (Google) Code Review <android-gerrit@google.com> | 2011-04-12 20:22:10 -0700 |
| commit | 47d2ef69d3f94e68bf675967ce3140cbb6179279 (patch) | |
| tree | 9b05e0415618003b0d64cc479e6e89726f134336 | |
| parent | 09df2a5bf4b15fdfc2e2033724f5cd926f94619d (diff) | |
| parent | dc5301e5902826b0e9a44fc3d45695526eaf66b3 (diff) | |
| download | latinime-47d2ef69d3f94e68bf675967ce3140cbb6179279.tar.gz latinime-47d2ef69d3f94e68bf675967ce3140cbb6179279.tar.xz latinime-47d2ef69d3f94e68bf675967ce3140cbb6179279.zip | |
Merge "Change the formula of the missing character."
| -rw-r--r-- | native/src/defines.h | 3 | ||||
| -rw-r--r-- | native/src/unigram_dictionary.cpp | 9 |
2 files changed, 8 insertions, 4 deletions
diff --git a/native/src/defines.h b/native/src/defines.h index 926120703..ff195f40c 100644 --- a/native/src/defines.h +++ b/native/src/defines.h @@ -138,7 +138,8 @@ static void prof_out(void) { #define SUGGEST_WORDS_WITH_SPACE_PROXIMITY true // The following "rate"s are used as a multiplier before dividing by 100, so they are in percent. -#define WORDS_WITH_MISSING_CHARACTER_DEMOTION_RATE 70 +#define WORDS_WITH_MISSING_CHARACTER_DEMOTION_RATE 90 +#define WORDS_WITH_MISSING_CHARACTER_DEMOTION_START_POS_10X 12 #define WORDS_WITH_MISSING_SPACE_CHARACTER_DEMOTION_RATE 80 #define WORDS_WITH_EXCESSIVE_CHARACTER_DEMOTION_RATE 75 #define WORDS_WITH_EXCESSIVE_CHARACTER_OUT_OF_PROXIMITY_DEMOTION_RATE 75 diff --git a/native/src/unigram_dictionary.cpp b/native/src/unigram_dictionary.cpp index c18829014..2ae2bca92 100644 --- a/native/src/unigram_dictionary.cpp +++ b/native/src/unigram_dictionary.cpp @@ -518,9 +518,12 @@ inline int UnigramDictionary::calculateFinalFreq(const int inputIndex, const int // TODO: Demote by edit distance int finalFreq = freq * matchWeight; if (skipPos >= 0) { - if (mInputLength >= 3) { - multiplyRate(WORDS_WITH_MISSING_CHARACTER_DEMOTION_RATE * - (mInputLength - 2) / (mInputLength - 1), &finalFreq); + if (mInputLength >= 2) { + const int demotionRate = WORDS_WITH_MISSING_CHARACTER_DEMOTION_RATE + * (10 * mInputLength - WORDS_WITH_MISSING_CHARACTER_DEMOTION_START_POS_10X) + / (10 * mInputLength + - WORDS_WITH_MISSING_CHARACTER_DEMOTION_START_POS_10X + 10); + multiplyRate(demotionRate, &finalFreq); } else { finalFreq = 0; } |