Final Year Report: Online Support Chatbot FYP[PDF]

10 Min Read

This report is the final year report of Rukesh Shrestha. Rukesh shrestha is graduate of university of Wolverhampton. We will provide you the link to his project very soon. You can download his report by clicking download here button at the end of this page.

1  Introduction

Services provide to the user/client by understanding the natural language is known as the chatbot. It is mainly used for the customer services. It has been used in different sector for the purpose of the information gathering or provide the required information and guidance. It is one of the AI program that simulates the human activities of the conversation or speech by using key pre-calculated user phrases, auditory and text based signals (Abdul-Kader & Woods, 2015). Artificial writing beings which can communicate to the human beings in certain bases such as text based, spoken conversation or the non-verbal conversation is the chatbot. It is the software or AI application which are develop to talk to human.  This application has become fascinating, stimulate and engrossing to entire world. 

History of Chatbot

Chatbot has the long history back to the mid-nineties. Lots of chatbot has been develop till now. Below is the time graph development of chatbot

Figure 1 Graphical Representation of chatbot history The detail of all history can be found in appendix 9.19.1. 

1.1       Problem Domain and Project as Solution 

Nowadays the internet has become basic needs of human beings. As it has become the basic need facility, to maintain the quality of connection all the time different ISP Companies have made different departments. Online Support is one of the important departments. Thousands of calls, e-mail messages, and chats are received due to the deterioration of the internet services provided by the ISP Company (Bergum & Iveland, 2019).All the calls and the messages sent by the user are not received. Lots of abundant calls and missed chats/messages are recorded. Manpower who is working in the concern department may suffer from different health problems as well as irritates from the chat. To maintain the quality of services chatbot can be an efficient software. It can solve the minor queries such as Wi-Fi password update, slow browsing, new connection queries, and update packages queries. Working eight hour a day by texting and talking to humans is impossible. Humans cannot be 100% efficient. While talking to clients, there may be some mistake. Same problem and question can be repeatedly asked by the client. Humans may get frustrated and lose their patience but the chatbot can answer the same question again and again without losing the patience and also encourages clients to solve the problem. In the long term view this technology can be proven as the cost effective related to human effort.

Type of Chatbot

There are mainly three types of chatbot. They are retrieval based chat-bot, generative based chat-bot and hybrid chatbot.  Chatbot contains a certain collection of the Frequent Ask Question (FAQ) template which match the keyword spoken by the user and generate the response is the retrieval based chatbot. Chatbot contains a neuron network and generates the reply from input data based on natural language generation techniques is generative based chatbot (Chowanda & Chowanda, 2018). This kind of chatbot is trained in the huge amount of real data. Combination of both retrieval and generative chatbot are the hybrid chatbot (IrynaKulatska, 2019).

In the final year project hybrid chatbot is going to be developed. This chatbot uses the Deep learning neural network model. Recurrent Neural Network (RNN) is used in this project. RNN is the power algorithm to address the problem of sequence data.  Long ShortTerm Memory (LSTM) architecture is used to develop the model.

LSTM is used for sequence to sequence problems as it has the cognition of memories the data input by the user and can easily train the huge sequence model architecture. The model is trained with the dataset and used to generate the responses (IrynaKulatska, 2019).

There are lots of applications of the LSTM. In this project, Conversational Modeling is preferred. The domain of this project is to generate the sequence of text from the asked question (Brownlee, 2017).

1.2       Academic Question

How could chatbot solve issue of the online support and how it works? 

1.3       Aims and Objectives

1.3.1     Aims

 To solve the relevant problems of online support prevailing in the ISP company by developing the chatbot.

1.3.2     Objectives

  • To develop the Hybrid chatbot.
  • To develop the chatbot using the neuron network.
  • To develop and deploy the chatbot in web application.
  • To maintain quality of services.
  • To solve the minor repetitive queries.

1.4       Artifact produced and background 

Artifact 1: Chatbot

It is a computer program which is developed to make the work comfortable and more accurate. Deep-learning component of AI is used to develop chatbot. Recurrent Neural Network (RNN) is going to be used for developing this chatbot. Long Short-Term Memory of RNN is going to be used for developing models. This model is mainly used in complex systems where the machine has to learn by understanding the human languages and posture. Keras sequential layers are used for retrieval based chatbot. 

Artifact 2: Web Application

In this project, a web application is developed to deploy the chatbot. Django framework is used to develop the web application. This framework is free and open source. It has its own server to deploy the web pages.


In the context of Nepal, there are many ISP Companies which provide chat services to clients. Thousands of queries are handled by the representative. Some of the queries are minor and some of them are complex to solve. Minor queries can be solved by the chatbot. Therefore, the representative feels less burden to work and tries to solve the complex issues.

1.5       Scope and Limitation of the project


Purpose system can guide clients to solve the minor issue such as updating the Wi-Fi password from mobile application and web application, troubleshoot the basic internet slow issue, and provide department number to client e. t. c.  Solving this issue will help to benefit employees as well as clients to not wait for a long time, each representative can handle the other issue at the same time. Limitation

 Purpose system is not able to solve the complex network issue such as:

o Latency issue face by the client. o Handle the physical network issue. o Check the client profile.

1.6       Introduce the structure of report


It includes the meaning of purpose system, its history, types, problem domain and project as the solution.

Literature Review

It includes the related project description, diagram and its results. Here, the literature review is divided in two part i.e. generative based review and the hybrid based review.

Supportive information

It includes the technical related information such as network definition and mathematics behind it. It includes the libraries and the framework description.

Main Body

It is divided in to two part i.e. AI model development and the software model development. Development includes who the system is developed and its phases with the development process.   

Answer to academic question

It includes answers to the academic question with the proof.


It includes what has been discovered while writing and developing the project.

Critical evaluation of the product

It includes the finding, difficulties and opinion.  

References and the Bibliography

It includes all the sources from where the project is developed and report is written.


It includes the project planning material, project visualization code, project code and some of the theoretical definitional.  

Share This Article
1 Comment

Leave a Reply

Your email address will not be published. Required fields are marked *