Haptik is India’s first Conversational Commerce platform that is powered by both AI and real humans and we aim to redefine the way people get their everyday jobs done using chat as the underlying interface.
To build conversational platforms like Haptik, one of the module that plays crucial role is Named Entity Recognition (NER). In this repository, we have open-sourced NER which uses heuristic approach along with several NLP techniques to extract necessary entities from chat interfaces.
To know more about our NER repository please go through the following documentations:
-
Installation and setup: Detailed document on how to install and setup the chatbot NER on your system
-
CSV file structure and adding/removing entities to DataStore: Document on how to create your own entity and store it to DataStore
-
Approach used: Document on the architecture along with different detection logics (entity types) and built-in entities
-
API calls: A brief document on the api calls for different detection logics along with its necessary output