Vocabulary Database for Reading Japanese (VDRJ)
Vocabulary Database for Reading Japanese (VDRJ) Ver. 1.1
Please download the database which meets with your needs.
Vocabulary Database for Reading Japanese (for Research) Ver. 1.1
Skewness, kurtosis of the sub-frequency distribution, sub-frequency rankings and its mean are added to Ver. 1.0, no other data correction is made.
The database which contains 141,950 types is provided as three separate database files. The extension of the databases is .xlsx.
Please open the files by Excel 2007 or a later version. The converter may be necessary with an older version.
VDRJ is made up of three databases: Top 60894, Assumed Known Words and Narrowly Ranging Words.
Please download the databases one by one from links below.
- 重要度順語彙データベース (Top 60894) 重要度順位 00001-60894 (42MB) VDRJ (Top60894) (42MB)
- 想定既知語彙データベース フィラー、記号その他、固有名詞 (16MB) VDRJ (Assumed Known Words: Proper Nouns, Fillers, Signs etc.) (16MB)
- 使用範囲狭小語彙データベース (24MB) VDRJ (Narrowly Ranging Words) (24MB) Download
Vocabulary Database for Reading Japanese (for Teachers) Ver. 1.0Download
The database is the simplified version of VDRJ for Research (Top60894). It would be enough for general educational purposes.
It contains the top 60894 words.
These are more simplified versions than the database for teachers;
however, the words are ordered not only for reading but also for daily life use. It contains approx. 20,000 words.
There are three types of databases: For General Learners, For International Students, and Basic 2500.
The first two includes explanations in simple Japanese.
Basic 2500, the database for beginners, has a simple English explanation.
The Vocabulary Database for Learners of Japanese Ver. 1.0 (for General Learners)Download
The Vocabulary Database for Learners of Japanese Ver. 1.0 (for International Students)Download
The Vocabulary Database for Learners of Japanese: Basic 2500, Ver. 1.0Download
Features of VDRJ：
The newest word frequency list made from the book corpus (approx. 28 million tokens) and internet forum corpus (5 million tokens).
--There have been some lists made from magazine or newspaper corpora but no list from a large book corpus.
--The list made from book corpora is better than the lists made from magazine or newspaper corpora in terms of the generality of the word-origin distribution and the stability of the word.
--The short coming from the book corpus is compensated by the internet forum vocabulary.
Dispersion is calculated from the sub frequencies in 10 sub corpora.
Words are ranked by the usage coefficient which is the product of frequency and dispersion.
--By taking the dispersion into account, unevenly distributed words are excluded from the high frequency band.
--Reordering words by applying different weighing to sub frequencies,
the new vocabulary ranking indices for general learners and international students are developed.
Adding word ranking for written Japanese together, three types of ranking are possible.
--Domain-specific words in the four science domains can be extracted
as the indices and levels for the domain-specificity of academic texts are added to the database.
--Possible literary words are extracted.