Entities are groups of keywords that share the same meaning, but can appear differently in every user’s query because of the user’s style of conversation, or because every user holds a unique value.
How is this useful?
You can build a dictionary of entities from keywords, nouns, and synonyms, so that the virtual assistant can identify them when placed in a user query.
- Entities teach your virtual assistant synonyms, short forms, service or product names and jargons
- Entities help virtual assistant detect specific user data or data formats with words to allow the virtual assistant to understand user intent eg. personal identification number, phone number, etc
- Entities accelerate virtual assistant training with less effort by combining intents that have similar variation
How does this work?
In order for the virtual assistant to understand jargon, abbreviations, particular nouns and synonyms, you must first let the virtual assistant know what these alternative ways of writing equate to. By telling the virtual assistant that these alternative words means the same thing, it will be better equipped in understanding the user’s intent.
For example:
Nouns
Master Card, Visa, AMEX = Credit Card entity
Fever, cough, blocked nose = Flu Symptoms entity
Synonyms/Jargon/Abbreviations
email, webmail, e-mail = Email entity
password, passwd, pw = Password entity
Similarly for regex strings, by adding them as entities, the Bot will be able to identify the pattern as an entity
For example:
Regex/Patterns
(0-1)/(a-z)@mail.com, (0-1)/(a-z)@gmail.com, (0-1)/(a-z)@hotmail.com, @yahoo.com = Email entity
SXXXXXX(A-Z) = NRIC/FIN entity
