Building an Enterprise Chatbot: Work with Protected Enterprise Data Using Open Source Frameworks - Couverture souple
Synopsis
Chapter 1: Processes in the Banking and Insurance Industry
The chapter will focus on explaining some core process within the banking and insurance industry that is suitable for a chatbot application.
No of pages: 30
Chapter 2: Identifying the Sources of Data
This chapter will discuss sources of data for conversation and action-based event triggers for a chatbot. Conversation courses would be from customer service centers, online chats, emails and other NLP sources, while action sources are customer account details and more personalize data.
No of pages: 30
Chapter 3: Mining Intents from the Data Sources
This chapter will discuss how to build a business-specific intent engine for chatbots.
No of pages: 30
Chapter 4: Building a Business Use-Case
This chapter will focus on how to identify the right business process to introduce chatbots. It will also discuss how to look at some of the metrics of success and RoI given a chatbot is deployed.
No of pages: 30
Chapter 5: Natural Language Processing (NLP)
Chapter Goal: This chapter focusses on processing and understanding natural language through the computer algorithm. It also introduces how to prepare data for applying the NLP algorithms. We will use Stanford CoreNLP, NLTK, gensim, OpenIE tools to explore and model.
No of pages: 80
Sub - topics
Introduction: Question & answering, information extraction, sentiment analysis, Machine translation,
Text processing: Regex, tokenization, normalization - lower case, lemmatization, stemming (Porters Algorithm), sentence segmentation
Converting text to features: Syntactical parsing - dependency grammar, PoS, entity parsing - phrase detection, topic modeling, statistical features - TF-IDF, word embeddings
Classification - spam filter using naïve Bayes, sentiment analysis using SVM on Lexicon and text feature.
NLP Tools - nltk, genism, openIE, CoreNLP
Chapter 6: Building Chatbots Using Popular Platforms
For general purpose chatbots, publicly available cloud services can be used to deploy chatbots faster and without any DevOps overhead. We shall discuss some of the major chatbot development platforms available in the market.
No of pages: 50
Sub-Topics
Microsoft Bot framework with LUIS
Google's DialogFlow
Amazon Lex with Lambda
Bottr, Chatfuel and others
Open framework RASA and Botpress
Chapter 7: Deployment and Continuous Improvement Framework
In this chapter we shall discuss and implement a custom built chatbot . We will discuss designing and implementing state machines and their different state transitions, and how they are critical to maintain the context of user utterance as well as in defining the chat flow using sessions that contains long term and short-term attributes.
No of pages: 50
Sub-topics:
Public endpoint creation
Intent engine development and deployment as API
Building state machine
Integration with Facebook messenger
Deployment of chatbot on AWS
Logging
Mining conversation log to improve intent engine
Recommending similar/next questions, pushing info
Les informations fournies dans la section « Synopsis » peuvent faire référence à une autre édition de ce titre.
