#30772We present a statistical model of Japanese unknown words consisting of a set of length and spelling models classified by the character types that constitute aword.
other,31-2-P99-1036,ak
ideograms
</term>
like
<term>
Chinese
</term>
(
<term>
kanji
</term>
) and
<term>
phonograms
</term>
like
<term>
#30805The point is quite simple: different character sets should be treated differently and the changes between character types are very important because Japanese script has both ideograms like Chinese (kanji) and phonograms like English (katakana).
other,34-2-P99-1036,ak
Chinese
</term>
(
<term>
kanji
</term>
) and
<term>
phonograms
</term>
like
<term>
English
</term>
(
<term>
katakana
#30808The point is quite simple: different character sets should be treated differently and the changes between character types are very important because Japanese script has both ideograms like Chinese (kanji) andphonograms like English (katakana).
measure(ment),5-3-P99-1036,ak
word segmentation accuracy
</term>
and
<term>
part of speech tagging accuracy
</term>
are improved by the proposed
<term>
#30820Both word segmentation accuracy andpart of speech tagging accuracy are improved by the proposed model.
measure(ment),1-3-P99-1036,ak
</term>
(
<term>
katakana
</term>
) . Both
<term>
word segmentation accuracy
</term>
and
<term>
part of speech tagging accuracy
#30816Bothword segmentation accuracy and part of speech tagging accuracy are improved by the proposed model.
model,15-3-P99-1036,ak
</term>
are improved by the proposed
<term>
model
</term>
. The
<term>
model
</term>
can achieve
#30830Both word segmentation accuracy and part of speech tagging accuracy are improved by the proposedmodel.
model,1-4-P99-1036,ak
the proposed
<term>
model
</term>
. The
<term>
model
</term>
can achieve 96.6 %
<term>
tagging accuracy
#30833Themodel can achieve 96.6% tagging accuracy if unknown words are correctly segmented.
other,6-1-P99-1036,ak
a
<term>
statistical model
</term>
of
<term>
Japanese unknown words
</term>
consisting of a set of
<term>
length
#30752We present a statistical model ofJapanese unknown words consisting of a set of length and spelling models classified by the character types that constitute a word.
other,17-2-P99-1036,ak
differently and the changes between
<term>
character types
</term>
are very important because
<term>
Japanese
#30791The point is quite simple: different character sets should be treated differently and the changes betweencharacter types are very important because Japanese script has both ideograms like Chinese (kanji) and phonograms like English (katakana).
other,29-2-P99-1036,ak
has both
<term>
ideograms
</term>
like
<term>
Chinese
</term>
(
<term>
kanji
</term>
) and
<term>
phonograms
#30803The point is quite simple: different character sets should be treated differently and the changes between character types are very important because Japanese script has both ideograms likeChinese (kanji) and phonograms like English (katakana).
other,21-1-P99-1036,ak
spelling models
</term>
classified by the
<term>
character types
</term>
that constitute a
<term>
word
</term>
#30767We present a statistical model of Japanese unknown words consisting of a set of length and spelling models classified by thecharacter types that constitute a word.
other,23-2-P99-1036,ak
types
</term>
are very important because
<term>
Japanese script
</term>
has both
<term>
ideograms
</term>
like
#30797The point is quite simple: different character sets should be treated differently and the changes between character types are very important becauseJapanese script has both ideograms like Chinese (kanji) and phonograms like English (katakana).
other,9-4-P99-1036,ak
96.6 %
<term>
tagging accuracy
</term>
if
<term>
unknown words
</term>
are correctly segmented . We propose
#30841The model can achieve 96.6% tagging accuracy ifunknown words are correctly segmented.
model,14-1-P99-1036,ak
words
</term>
consisting of a set of
<term>
length and spelling models
</term>
classified by the
<term>
character
#30760We present a statistical model of Japanese unknown words consisting of a set oflength and spelling models classified by the character types that constitute a word.
model,3-1-P99-1036,ak
Construct Algebra
</term>
. We present a
<term>
statistical model
</term>
of
<term>
Japanese unknown words
</term>
#30749We present astatistical model of Japanese unknown words consisting of a set of length and spelling models classified by the character types that constitute a word.
other,36-2-P99-1036,ak
</term>
) and
<term>
phonograms
</term>
like
<term>
English
</term>
(
<term>
katakana
</term>
) . Both
<term>
#30810The point is quite simple: different character sets should be treated differently and the changes between character types are very important because Japanese script has both ideograms like Chinese (kanji) and phonograms likeEnglish (katakana).
other,7-2-P99-1036,ak
point is quite simple : different
<term>
character sets
</term>
should be treated differently and
#30781The point is quite simple: differentcharacter sets should be treated differently and the changes between character types are very important because Japanese script has both ideograms like Chinese (kanji) and phonograms like English (katakana).
measure(ment),6-4-P99-1036,ak
<term>
model
</term>
can achieve 96.6 %
<term>
tagging accuracy
</term>
if
<term>
unknown words
</term>
are correctly
#30838The model can achieve 96.6%tagging accuracy if unknown words are correctly segmented.
other,38-2-P99-1036,ak
phonograms
</term>
like
<term>
English
</term>
(
<term>
katakana
</term>
) . Both
<term>
word segmentation accuracy
#30812The point is quite simple: different character sets should be treated differently and the changes between character types are very important because Japanese script has both ideograms like Chinese (kanji) and phonograms like English (katakana).