大会名称 |
---|
2018年 情報科学技術フォーラム(FIT) |
大会コ-ド |
F |
開催年 |
2018 |
発行日 |
2018-09-12 |
セッション番号 |
4p |
セッション名 |
人文科学と社会基盤 |
講演日 |
2018/09/20 |
講演場所(会議室等) |
E棟Cul Site R1 |
講演番号 |
N-012 |
タイトル |
Inferring CEFR Reading Comprehension Index Based on Japanese Document Classification Method Including Pre-A1 Level |
著者名 |
Nguyen Tra My Huynh, Yoshinori Miyazaki, Seiji Tani, |
キーワード |
CEFR, Classification, Machine Learning, Corpus, Can-Do Statement |
抄録 |
The CEFR (Common European Framework of Reference for Languages) Companion Volume is intended as a complement to the CEFR published in 2001. This study will continually work on the CEFR for Japanese learners, adding new seven Can-Do Statements (CDSs for short, as CEFR reading comprehension indices) of Pre-A1 level in the Companion Volume, and carrying out experiments for the classification of Japanese documents into their corresponding CDSs. From three former features (length, document type and technicality), we have added a new one: Kanji rate considered as one of the vital elements in reading comprehension. Further, instead of conventional 7 document types, we divided sentences into 8 types. The results of experiments will be shown in the presentation. |
本文pdf |
PDF download (295.2KB) |