REACT/NODE/MONGO CHATBOT
Budget: £250 – £750 GBP
Overview
The intention of the React/Node/Mongo Chatbot Project Is to build a chatbot using a NERM stack and Javascript libraries such as NLP.js, Promise, Natural and Jovo.
Purpose
The purpose of React/Node/Mongo ChatbotIs to provide a user with a structured conversational process. Unlike most chatbots that are designed to answer questions, this chatbot will instead ask the user questions.
The chatbot must be able to recognise conventional conversational utterances and intents such as greetings and goodbyes but the main part of the conversation will be driven through the chatbot recognising custom named entities in the user’s answers to the chatbot’s questions.
The custom named entities are currently stored in a MySQL database and in this project they will be read from a MongoDB database or a json file if that does not impact performance too heavily There are currently around 12,000 of these custom named entities.
The chatbot operates by starting off with greeting items and entities in the way a normal chatbot does and then asks the user a question or questions from which it can recognise custom named entities. Using the MongoDB or json file as a lookup table, the chatbot then selects questions from a database and asks the user these questions. As the user responds and the conversation continues the chatbot will identify more custom named entities from a database and ask further questions based on these.
The conversation with the user is structured in stages, so the chatbot will ask different questions about the custom named entities being shared as the conversation progresses. Towards the end of the conversation, the chatbot will ask the user what action they need to take and then the conversation will finish with the conventional type of chat bot goodbyes and cheerios.
A transcript of the conversation will be stored in a database for future reference. The ability to use ASR and TTS with a library such as Jovo will be welcomed.
Results
The result of the project will be a chatbot that can generate a conversation with a user by asking a structured series of questions based on the recognition of custom named entities. Transcripts of the conversations will be stored and the capacity to use a voice channel be welcomed.
The intention of the React/Node/Mongo Chatbot Project Is to build a chatbot using a NERM stack and Javascript libraries such as NLP.js, Promise, Natural and Jovo.
Purpose
The purpose of React/Node/Mongo ChatbotIs to provide a user with a structured conversational process. Unlike most chatbots that are designed to answer questions, this chatbot will instead ask the user questions.
The chatbot must be able to recognise conventional conversational utterances and intents such as greetings and goodbyes but the main part of the conversation will be driven through the chatbot recognising custom named entities in the user’s answers to the chatbot’s questions.
The custom named entities are currently stored in a MySQL database and in this project they will be read from a MongoDB database or a json file if that does not impact performance too heavily There are currently around 12,000 of these custom named entities.
The chatbot operates by starting off with greeting items and entities in the way a normal chatbot does and then asks the user a question or questions from which it can recognise custom named entities. Using the MongoDB or json file as a lookup table, the chatbot then selects questions from a database and asks the user these questions. As the user responds and the conversation continues the chatbot will identify more custom named entities from a database and ask further questions based on these.
The conversation with the user is structured in stages, so the chatbot will ask different questions about the custom named entities being shared as the conversation progresses. Towards the end of the conversation, the chatbot will ask the user what action they need to take and then the conversation will finish with the conventional type of chat bot goodbyes and cheerios.
A transcript of the conversation will be stored in a database for future reference. The ability to use ASR and TTS with a library such as Jovo will be welcomed.
Results
The result of the project will be a chatbot that can generate a conversation with a user by asking a structured series of questions based on the recognition of custom named entities. Transcripts of the conversations will be stored and the capacity to use a voice channel be welcomed.