E-048 ウイグル語単音節の出現頻度とその特徴分析(自然言語・音声・音楽,一般論文)

概要

論文の詳細を見る
Although Uyghur language is being spoken by more than ten million of people in central Asia, speech typewriter for this language has not been developed. In this paper, we analyzed syllables frequency and characteristics of Uyghur language to find the significant syllables to create an efficient database for future speech typewriter. The analyses were based on a modern Uyghur word dictionary and conversational sentences selected arbitrarily from Uyghur websites. 37,255 unique words in Uyghur modern Dictionary and 30,439 in conversational sentences were syllabified manually, resulting 2,557 and 1,437 syllables, respectively. 1,390 of these syllables share Saimaiti and Feng's (2007) which can be considered as most frequently used syllables. This number covers 98.14% of the entire Uyghur syllables. The problem arises on creating speech typewriter because of too many syllables. However, using only the 500 most frequent syllables which cover more than 90% of the entire Uyghur syllables, the speech typewriter of Uyghur language can be created efficiently.
FIT(電子情報通信学会・情報処理学会)推進委員会の論文
2009-08-20