Skip to content

tw1270/Chatbot

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Why your business needs a Chatbot?

Tenzin Wangdu

May 03, 2021

Table of Contents


Background


  • First Chatbot was developed in 1966 at MIT called ELIZA
  • ELIZA was simple decision tree questions that answer a few questions
  • Now it is developed into everyday life with messenger apps, voice assistant
  • Chatbots are quickly replacing human for technical support and customer service

Problem Statement


Every business needs a chatbot for their website or app. Chatbot can replace a customer service agents for a 24 hour services and help business save money

Different methods


Creating your own data/intents

  • Intents are categories of the text of user’s input
  • Ex: ‘Hi’ would be a greeting, ‘how can you help me?’ would be a help intents
  • Creating different intent for different purpose
  • It is Great for FAQ and easy to create
  • Creating different responses to those intents

RASA Framework

  • Creating Intents
  • Allows us create a storyline and it can store the previous answers
  • So, the bot can continue a conversation

RASA Pipeline

  • Whitespace Tokenizer (using whitespaces as a separator)
  • Count Vectors Featurize(Creates bag-of-words representation of user messages, intents, and responses)
  • N-gram from 1-4
  • NLU model (Natural language understanding)

Creating your own framework

  • Data: Twitter Customer Service Tweets(3.8 millions tweets)
EDA/Preprocessing the Customer Tweet
  • Remove all the non-english tweets
  • Lemmatize
  • Remove stopwords, href, @ handles
  • Setting a limit on the length of the tweet at 5-40
Creating Intents
  • The interpretation of a statement is what allows chatbot to formulate the best possible response.
  • Matching tweets with the intents of the customer (Battery, Update, Macbook, and etc)
Modeling
  • RNN Neural Network to train the model
  • It had a 99.4% of train accuracy, and 93% of testing accuracy

Overall Conclusions


  • Chatbot allows business to be to available to customer 24x7
  • Huge expense cut/Alternative to customer service if needed
  • Building chatbot based on your business

Areas for Further Research


  • Launching my own framework
  • Deploying the other model on flask and heroku
  • Connected the chatbot to SQL

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published