| measure(ment),1-3-P99-1036,ak | </term> ( <term> katakana </term> ) . Both <term> | word segmentation accuracy | </term> and <term> part of speech tagging accuracy | #30816 Bothword segmentation accuracy and part of speech tagging accuracy are improved by the proposed model. | |
| 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> | #30820 Both word segmentation accuracy andpart of speech tagging accuracy are improved by the proposed model. | |
| 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 | #30838 The model can achieve 96.6%tagging accuracy if unknown words are correctly segmented. | |
| model,1-4-P99-1036,ak | the proposed <term> model </term> . The <term> | model | </term> can achieve 96.6 % <term> tagging accuracy | #30833 Themodel can achieve 96.6% tagging accuracy if unknown 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 | #30760 We 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,15-3-P99-1036,ak | </term> are improved by the proposed <term> | model | </term> . The <term> model </term> can achieve | #30830 Both word segmentation accuracy and part of speech tagging accuracy are improved by the proposedmodel. | |
| model,3-1-P99-1036,ak | Construct Algebra </term> . We present a <term> | statistical model | </term> of <term> Japanese unknown words </term> | #30749 We 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,17-2-P99-1036,ak | differently and the changes between <term> | character types | </term> are very important because <term> Japanese | #30791 The 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,21-1-P99-1036,ak | spelling models </term> classified by the <term> | character types | </term> that constitute a <term> word </term> | #30767 We 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 | #30797 The 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,26-1-P99-1036,ak | character types </term> that constitute a <term> | word | </term> . The point is quite simple : different | #30772 We 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,27-2-P99-1036,ak | <term> Japanese script </term> has both <term> | ideograms | </term> like <term> Chinese </term> ( <term> kanji | #30801 The point is quite simple: different character sets should be treated differently and the changes between character types are very important because Japanese script has bothideograms 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 | #30803 The 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,31-2-P99-1036,ak | ideograms </term> like <term> Chinese </term> ( <term> | kanji | </term> ) and <term> phonograms </term> like <term> | #30805 The 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 | #30808 The 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). | |
| other,36-2-P99-1036,ak | </term> ) and <term> phonograms </term> like <term> | English | </term> ( <term> katakana </term> ) . Both <term> | #30810 The 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,38-2-P99-1036,ak | phonograms </term> like <term> English </term> ( <term> | katakana | </term> ) . Both <term> word segmentation accuracy | #30812 The 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,6-1-P99-1036,ak | a <term> statistical model </term> of <term> | Japanese unknown words | </term> consisting of a set of <term> length | #30752 We 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,7-2-P99-1036,ak | point is quite simple : different <term> | character sets | </term> should be treated differently and | #30781 The 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). | |
| other,9-4-P99-1036,ak | 96.6 % <term> tagging accuracy </term> if <term> | unknown words | </term> are correctly segmented . We propose | #30841 The model can achieve 96.6% tagging accuracy ifunknown words are correctly segmented. |
