TABLE OF CONTENTS

TITLE PAGE

CERTIFICATION

DEDICATION

ACKNOWLEDGEMENTS

ABSTRACT

TABLE OF CONTENTS

LIST OF FIGURESix

LIST OF TABLES

CHAPTER ONE: INTRODUCTION

1.1 Background to the Study 1

1.2 Statement of Problem 3

1.3 Aim and Objectives 3

1.4 Significance of the Study 4

1.5 Scope of the Study 4

1.6 Definition of Terms 5

1.7 Project Layout 6

CHAPTER TWO: LITERATURE REVIEW

2.1 Chatbots 7

2.2 Types of Chatbots 9

2.3 How Do Chatbots Work? 10

2.3.1Options to Build a Chatbot 10

2.3.2Chatbot Platforms Alternatives 11

2.4 Design Principles 12

2.4.1Rule-Based System 12

2.4.2Machine Learning System 14

2.5 Artificial Intelligence 14

2.5.1Artificial Neural Networks 14

2.5.2Machine Learning 15
ii
 2.5.3Natural Language Processing (NLP) 16

2.5.4Cognitive Computing 16

2.5.5Firebase Database 16

2.5.6Google Dialogflow 17

2.5.7Android Mobile Operating System 17

2.6 Chatbot Engineering and Design Approaches 18

2.7 Common Technologies Used in Chatbots 19

2.8 Data Management and Training Methods for Dialogue Systems 20

2.8.1Training Dialogue Systems 25

2.9 Contextuality in Intent-Based Chatbots 29

2.10 A Concept of Continual Learning 32

2.11 Concerns about Catastrophic Forgetting 35

2.12 Chatbot Application on Websites 36

CHAPTER THREE: RESEARCH METHODOLOGY

3.1 Overview 37

3.2 System Architecture 37

3.3 PyTorch Modelling 40

3.4 Python Programming 40

3.5 Web Development 40

CHAPTER FOUR: IMPLEMENTATION, RESULTS AND ANALYSIS

4.1 Implementation 41

4.2 Software prerequisites for running the case study chatbot 41

4.3 Initial preparations before the implementation 41

4.4 Intent-based chatbot implementation 42

4.5 Module for continual learning 53

4.6 Designing the web interface 57

4.7 Results 63

4.8 Analysis of Results 68

4.9 Cost/Time Analysis 68

iii
 4.10 Project Limitations 69

CHAPTER FIVE: DISCUSSION, CONCLUSION AND RECOMMENDATIONS

5.1 Discussions 70

5.2 Recommendations 71

5.3 Conclusion 71

REFERENCES 73

LIST OF FIGURES

Figures Pages

2.1: Steps of an interaction between a user and a continual learning chatbot 34

3.1: Data flow diagram 38
3.2: User Roles model 39
4.1: Preview of the JSON file containing user intents and bot’s responses 44
4.2: PyTorch model stemming the “bag of words” 46
4.3: Snapshot of the number of unique words at the end of stemming 47
4.4: Looping through each sentence of the intent’s patterns 48
4.5: PyTorch populating the empty arrays with tokens and intents 49
4.6: Implementation of the bag words 51
4.7: Training the model 52
4.8: Defining the method to call up the chat module 54
4.9: Calling up the chat module in the command line 55
4.10: HTML being used to design the main interface of the web interface 58
4.11: An overview of the CSS code used in designing the web interlace 59
4.12: Demonstration of the JavaScript code used in designing the web interface 60
4.13: Landing page of the chat module 61
4.14: The chat module in use 62
4.15: The questionnaire used to determine the system’s efficiency 64
4.16: Pie chart displaying the relevance of the bot’s responses 65
4.17: Pie chart displaying the speed of the bot’s responses 66
4.18: Pie chart displaying the satisfaction of the users 67

iv
4.13: Landing page of the chat module
 LIST OF TABLES

Tables Pages

4.1: Version of used software and libraries ` 43

CHAPTERONE

INTRODUCTION

1.1 Background to the Study

In this technology-driven world, access to information has been made easier than ever. Continuous

modifications in the interactions among individuals daily, prompted by technological progress, have

significantly enhanced operational and professional efficiency. Conversation is an interesting type of

interaction, solely because it bridges the gap of communication. It can occur in human-human interaction

(HH) and human-computer interaction (HCI).

A chatbot is a tool that can simulate this kind of interaction byunderstanding the context and deliveranceof

an appropriate response.It is a program that simulates a conversation between a human and a computer.

Chatbots are used a lot in customer interactions, marketing on social network sites, and instant messaging

the client. Integrating chatbots into websites can be very profitable for business owners who are trying to

reach a new audience. (Macdonald, 2019)

Entrepreneurs have realized that without a fast, multilingual customer-service team that offers 24/7 support

to the customers, it will be difficult to succeed in any industry. With the integration of a chatbot intoa

website, this service is provided at the customer’s beck and call. While chatbots cannot hold a truly natural

conversation, they perform well at being consistent as they will only do what they are programmed to do.

Strutynskyi(2020) stated that chatbots cannot replace humans in sales completely, but they can take over

answering repetitive questions and taking simple orders. For cases when advanced support is needed, a “Call

a human agent” button is often present. With this short overview done, it is safe to say that since their first

appearance, chatbots have been bringing revolutionary changes to the business world and will keep doing

so.

v
Huang (2021) stated that artificial Intelligent machines can operate tasks of varying complexity, from

general work to sophisticated operations such as object detection using computer vision. Prominent trends in

this field are human brain simulation, natural-language processing and neural networking to name the least.

An example of an Artificial Intelligence system is a Chatbot. A chatbot is an intelligent conversational

computer program that mimics human conversation in its natural form using interactive textual or speech

skills.

Since the introduction of the first of its kind, ELIZA, in 1966, chatbots have evolved with the progressive

increase in computational capabilities and advances in Natural Language Processing tools and techniques.

ELIZA was a landmark system that inspired further research in the field. The application of chatbots has

expanded with the emergence of Deep Learning algorithms. One of the new, and the most interesting

application, is the development of smart personal assistants (such as Amazon’s Alexa, Apple’s Siri, Google’s

Google Assistant (Google Developers, 2023), Microsoft’s Cortana, and IBM’s Watson).

Voice assistants such as Apple SIRI, Microsoft Cortana, Amazon Alexa, and Google Assistant are widely

spread in today’s world. These systems are continually developing thanks to breakthroughs in speech

recognition and text-to-speech technologies. A major role in this development plays new research in the area

of deep learning, gains in computing power of GPUs and releases of ever more powerful mobile devices.

A chatbot can process user input and produce an output. Usually, chatbots take natural language text as

input, and the output should be the most relevant response to the user input sentence.Chatbots can directly

converse with users using artificial intelligence and Machine Learning concepts. It can also be defined as an

“online human-computer dialogue system(s) with natural language”. Chatbots constitute therefore an

automated dialogue system, that can attend to thousands of potential users at once.Usually, a chatbot

program understands one or more human languages although a general understanding of naturally-spoken

language across multiple dialogue steps, even in single task-oriented situations still cannot be achieved.

Hence, the system interprets human language input using information fed to it from the model or the

database. Creating a chatbot system for websites enables quick question response and problem-solving. This

servesa great purpose in saving a lot of money and time.

vi
1.2 Statement of Problem

Website owners face a lot of issues ranging from providing customer support to engaging and interacting

with users at various points in time. Hence, businesses waste a lot of capital and time, thereby, reducing their

efficiency in the business sector. There is so much a human can do when providing help. This is where a

web-based chatbot system comes in.(Huang, 2021)

The chatbot system aims to replace the human personnel that should be doing this job. The system will

constitute a chat CUI (Conversational User Interface) that will cater for different and varying needs per user.

The system consists of a chat module that interacts with the user’s input. The chatbot system can help

customers find the right services and products as well as answer their questions.

1.3 Aim and Objectives

Thisaim of this research is to design and implement a web-based chatbot system using Python and NLTK

(Natural Language Toolkit) and the specific objectives are:

(i) to process extracted data into a model using PyTorch;

(ii) to design and implement a web-based chatbot system to interact with website users and auto-

respondto user questions;

(iii) to design a descriptive land-on webpage to host the chatbot system;

(iv) to analyse the results of the chatbot and test if it is functional.

1.4 Significance of the Study

The use of web-based chatbots comes in handy in this generation due to easy access to mobile devices and

the internet. It is a technological revolution that is impacting the user experience of many websites as it

helps with easy communication. The use of AI and chatbots improve the quality of day-to-day

operations.Chatbots are developed and become so popular due to the increased use of smart devices and IoT

technology. MacDonald (2019) There have been several attempts at solving the problem of expanding

datasets. Some authors suggest using a NLP library with a programming language like Python, some also

suggest using data pre-sets to determine what response the chatbot system gives. In this work,a combination
vii
of datasets with NLP librariesis used to implement a chatbot system. Some research has been done in this

area to define and prove the effectiveness of this concept and the exploration of this concept is currently

ongoing. Strutynskyi (2020) Despite the amount of existing work in the area of machine learning, this

concept has not yet been investigated concerning chatbots, which is the topic of this work.

A more detailed literature overview is presented in the Literature review section of this thesis.

1.5 Scope of the Study

(deBayser, 2017) This work is focused majorly on helping out people who are trying to integrate a round-

the-clock chatbot system in their websites, but are struggling to do so. It might also be interesting for

researchers working in the area of Machine Learning, companies looking for a suitable chatbot for business

applications, or students seeking further knowledge about different areas of AI.

1.6 Definition of Terms

Accuracy: The degree to which the result of a measurement, calculation and specification conforms to the

correct value or standard.

Artificial Intelligence (AI): The simulation of human intelligence processes by machines, especially

computer systems. Specific applications of AI include expert systems, natural language processing, speech

recognition and machine vision.

Conversational User Interface (CUI): A user interface for computers that emulates a conversation with a

real human.

Data: Information in its raw form must be processed before it can become meaningful information.

Dataset: A collection of related data or information composed of separate elements which can be

manipulated as a unit by a computer.

Graphical User Interface (GUI): a form of user interface that allows users to interact with electronic

devices through graphical icons and audio indicators such as primary notation, instead of text-based UIs,

typed command labels or text navigation.

Machine Learning (ML): An AI technique that teaches computers to learn from experience.
viii
Natural Language Processing (NLP): The application of computational techniques to analyse and

synthesise natural language and speech.

Natural Language Understanding (NLU): A subtopic of natural-language processing in artificial

intelligence that deals with machine reading comprehension.

Natural Language Toolkit (NLTK): A suite of libraries and programs for symbolic and statistical natural

language processing for English written in Python.

1.7 Project Layout

This project is organized as follows:

Chapter One: presents the background to the study, statement of the problem, aim and objectives of the

study, significance of the study, definition of terms and project layout.

Chapter Two: contains the overview of existing work in the area of building chatbots, training them using a

dialogue corpus and a description of the related concepts

Chapter Three:discusses the severaltechniques employed in achieving the different objectives of the study.

Chapter Four:presents the results and discussion.

Chapter Five:entails the summary, conclusion and recommendations.

ix