Introduce 係り受け解析 (Dependency Analysis) - based Japanese reading ease scores
kennysong opened this issue · comments
In the survey of Japanese reading ease scores, there are more complicated metrics that require 係り受け解析 (Dependency Analysis). Using multiple scores to measure the reading ease from multiple dimensions is recommended in the paper, so we should implement these metrics, too.
These are the libraries we could use:
https://github.com/megagonlabs/ginza
https://github.com/ku-nlp/knp/
https://taku910.github.io/cabocha/
The modules could be used for input augmentation on the Japanse texts.