Problem extracting the similar entities

i am building bot using the Restaurant Use case which has been considered while academy training . I have 2 queries on the same.

a. if we are creating a guided flow and giving user quick replies to select an option but user instead of choosing an option ,writes down his own query then my bot is not responding anything as its not been able to understand what to do …I want the bot to understand and extract entity and intent and should go to the particular step in the journey

b. if a user gives an utterance like i want to order dal and dal as a category belongs to both available cuisines (north as well as south ) then how would you train your bot in such a way that it should able to understand which cuisine we are talking and which dish we are talking about

Entities
1.cuisine type - north indian ,south indian
north indian has synonyms -kulcha,chole bhature ,north
south indian synonyms- dosa,idly,south

  1. fooditem-to get the food item name

Could anyone please guide on the same ?

1 Like

Hi Vani,

Here are the observations to your queries:

a. If a user does not select any of the buttons in quick replies, the fallback branch in the quick replies should handle the custom responses which are not part of the buttons.
Even if a user types in a response to the quick reply, as long as it contains trained entities and the entities are mapped to a node which is to be skipped, the bot can take the user to the intended node directly.

b. When training the bot on understanding dal as an entity for both north and south indian, give different synonyms in order to avoid conflicts. If conflicts arise from a user utterance, the NLP prediction model will determine which item is displayed