Requesting to allow users to define their list of user_dictionary words for the purpose of Kuromoji’s tokenizer.
I understand that WDS uses Kuromoji underneath for Japanese language. A key function of Kuromoji as the morphological analyzer for Japanese text is its ability to tokenize words (because in Japanese there is no spacing to signal separation between words like other languages such as in English). To ensure this capability works for a user’s domain Kuromoji comes with the ability to append a “user_dictionary” in addition to its default list, it is this capability that I am requesting be exposed for WDS users in Japanese.
This function is considered a necessity for Japanese large clients, including our current largest user of R&R at JR-East. The reason being, in many client’s call center support operations, including JR-East’s, they contain business terminology that’s unique to their business. These business terminology are words created from multiple nouns joined together (i.e compounds of multiple nouns). Whether these business terms are treated (tokenized) as a single compound and/or as multiple separate nouns, greatly impacts the accuracy of searches.
Why is it useful?
|Who would benefit from this IDEA?|
How should it work?