Unnecessary space in analyzed entity value
When LUIS takes more than 2 words in an entity, it returns result with unnecessary blank (in our language). It happens even trained without blank.
It happens in Japanese, and it's assumed to happen to other languages too (Chinese, Korean,etc).
Here attached sample case,
- trained app without blank (BurgerShopBot201906)
- got response (BurgerShopBot201906_response)
"チーズバーガー" (cheese burger) is analyzed as "チーズ_バーガー" with unnecessary blank, which is assumed as LUIS analyzed 2 words : "チーズ"(cheese) and "バーガー"(burger). This app is trained "チーズバーガー" as an entity without blank (no word breaker blank).
We use our local language (Japanese) we don't use blank as word breaker, and sometimes mix English and/or other languages which need blank as word breaker.
That means we can't deal LUIS adds unnecessary blank since we have both case that needs blank (when mixing)