‫بسم الله الرحمن الرحيم‬

Sudan Universality of Science and Technology

Collage Of Computer Science and Technology
Department Of Computer System and Network

Graduation project
Admission Enquiry Chat-Bot

Prepare by :
Elaf Dafallah Abdalla
Zakiay Hassan Abdalla Eissa

Supervisor :
Intissar

Date :

1
 Abstract

Today's widespread use of smartphones is proof that technology is always evolving.

Nowadays, artificial intelligence is crucial to numerous industries, including
manufacturing, human resources, and customer service. There are numerous chatbots
that help people discover solutions to their questions. As a result, we are developing an
AI-powered chatbot that can address all questions about colleges. It serves as an
intelligence tool with an emphasis on higher education. This artificially intelligent machine
will respond to queries from users regarding matters related to higher education. The
information is kept in the chatbot's database so that it can be used to spot trends and
decide how to respond to questions. The chatbot for college inquiries was developed
using an NLP system that evaluates questions and comprehends messages. People react to
each other differently based on emotions and attitude. Because chatbots must go by rules
just like humans do, they will communicate with customers in a courteous and correct
manner. Students can ask the chatbot any questions at any time of the day or night, and
they will receive a prompt and accurate response. A chatbot can respond to thousands of
users simultaneously. Chatbot can work 24 /7 without getting tired. It has minimal errors,
which increases productivity

2
 Table of content

Abstract....................................................................................................................................................................2

Chapter 1: Introduction..................................................................................................................................................5

1.1 Context...............................................................................................................................................................5

1.2 Statement of The Problem.................................................................................................................................5

1.3 Aims and Objectives...........................................................................................................................................6

1.4 Achievements.....................................................................................................................................................7

1.4 Methodology.....................................................................................................................................................8

1.6 Contributions.....................................................................................................................................................8

Chapter 2 Literature Review.........................................................................................................................................10

2.1 chat bot:...........................................................................................................................................................10

2.2 Learning chatbots............................................................................................................................................10

3.2 Sudanese Universities Using Chat-bots :.........................................................................................................11

2.4 Development of Chatbot-Provided Registration Information Services for Distance Learning Students:......11

2. 5 Prototyping a Chatbot for Student Supervision in a Pre-Registration Process:.............................................13

2.6 Chatbots in Education and Research: Ethical Implications and Solutions:.....................................................13

2.7 College Enquiry Chatbots Using Iterative Model:............................................................................................14

Chapter 3: Analysis and design......................................................................................................................................16

3.1 Introduction.....................................................................................................................................................16

3.2 Functional Requirements :...............................................................................................................................16

3.3 Non-Functional Requirements.........................................................................................................................17

3.4 System Components:......................................................................................................................................19

3.5 Tools and Technologies....................................................................................................................................20

3.6 Natural Language Processing (NLP) ................................................................................................................20

3 . 7Use case model..............................................................................................................................................24

Chapter 4: Design..........................................................................................................................................................26
4.1 Introduction:.....................................................................................................................................................26
4.2 Setting up the Environment and Installing Tools.............................................................................................26
4.3 Creating a Rasa Project.........................................................................................................................27

3
4.4 Designing and Integrating the User Interface with Rasa.................................................................................28

4
Chapter 1: Introduction

1.1 Context

The process of university admissions is often complex and time-consuming for both
prospective students and administrative staff. With the increasing number of applicants,
universities face significant challenges in managing inquiries related to application forms,
eligibility, deadlines, scholarships, and program details. Traditional communication
channels, such as phone calls and emails, tend to overload the admissions offices,
resulting in delayed responses and administrative inefficiency.
In recent years, advancements in artificial intelligence (AI) and natural language processing
(NLP) have provided innovative solutions to automate such processes. Chatbots, powered
by AI, have been widely adopted in various sectors, including customer service,
healthcare, and education, to enhance user experience and operational efficiency. In the
context of higher education, AI-powered chatbots are particularly useful in automating
responses to routine admission inquiries. This research focuses on developing an
Admission Enquiry Chatbot, which is designed to automate the university admission
process, providing prospective students with quick and accurate responses to their
inquiries, thereby easing the workload on administrative staff.

1.2 Statement of The Problem

The Admission Enquiry Chatbot project addresses several challenges that universities face
during the admission process, particularly during peak application seasons. One of the key
problems is the overwhelming volume of inquiries that universities typically receive.
During peak periods, prospective students flood the university’s admissions office with
questions about deadlines, programs, scholarships, and eligibility criteria. Handling this
volume of inquiries manually can strain staff, leading to delays in responses and potential
miscommunications.
Another challenge is the inefficiency of traditional communication channels. Whether via
email or phone, these methods can be slow, prone to errors, and require students to wait
for replies, which can cause frustration and

confusion. Students are often left uncertain about where they stand in the admission
process, especially when response times are lengthy.
5
Additionally, limited accessibility is a significant issue. Universities generally operate
within standard working hours, making it difficult for students in different time zones or

those with busy schedules to access the information they need. This lack of timely access
to essential admission details could lead to students missing deadlines or missing out on
critical information.
The chatbot aims to solve these problems by offering an instantaneous, 24/7 response
system. It will handle the repetitive questions, offer real-time assistance, and ensure that
students receive accurate and timely information, anytime, anywhere. Furthermore, the
chatbot can personalize interactions, making the process more relevant to each
applicant's needs, whether it's guiding them through the application process or providing
detailed answers based on their academic interests.

1.3 Aims and Objectives

The primary aim of this research is to design and develop an AI-powered Admission
Enquiry Chatbot that can automate and streamline the university admission process. The
specific objectives of this research are as follows:
1. Development of the Chatbot: To create a chatbot that leverages Natural Language
Processing (NLP) and machine learning techniques to understand and process user
queries related to university admissions.
2. Integration with Databases: To integrate the chatbot with a database that stores
relevant admission information such as application deadlines, eligibility criteria,
scholarship details, and program requirements.
3. User Interface Design: To design a simple, intuitive, and user-friendly web interface
that allows prospective students to interact with the chatbot and submit their
queries easily.
4. System Evaluation: To test the chatbot's functionality by collecting feedback from
prospective students and evaluating its effectiveness in providing accurate, timely,
and relevant information.
5. Improvement and Optimization: To continuously improve the chatbot based on
testing results, including adjusting its NLP model and training it to handle diverse
user queries more effectively.

6
By achieving these objectives, the chatbot will aim to enhance the accessibility and
efficiency of the admission process, enabling 24/7 assistance for prospective students and
reducing the administrative burden on university staff.

1.4 Achievements
As part of this project, a working prototype of the Admission Enquiry Chatbot has been
developed, which is capable of answering basic queries related to university admissions.
The chatbot has been designed using the Rasa framework, which facilitates the
implementation of Natural Language Understanding (NLU) and dialogue management. Key
achievements of this research include:
 Chatbot Development: The chatbot successfully processes text-based queries,
interprets user intent, and provides automated responses based on the information
stored in the database.
 Database Integration: A MySQL database has been implemented to store
admission-related data, such as deadlines, eligibility criteria, and program
information, enabling the chatbot to provide accurate and up-to-date responses.
 User Interface: A web-based interface has been developed, allowing students to
interact with the chatbot in a simple and efficient manner.
 User Testing: Initial user testing has been conducted with a group of prospective
students, gathering feedback on the chatbot’s usability, accuracy, and response
time. This feedback will be used to improve the chatbot in subsequent development
phases.
These accomplishments have paved the way for further refinement and deployment of
the chatbot, ultimately contributing to the improvement of the university’s admissions
process.
This iterative development process, along with regular testing and feedback, ensured that
the chatbot evolved in line with the project’s objectives and user requirements.

1.4 Methodology
The design and development of the system followed the agile model as described below.

7
 Figure 1.1 aglie model

Agile model in the project Chatbot for expertise inquiries to better participate in the
development of the system via short sprints. Allows continuous subscriptions from users,
enabling the activation of the system. Continuous engagement with student jobs allows
for quick modification of tasks on their items, increasing and ensuring that actual value is
delivered with all releases.

1.6 Contributions
This research contributes to the field of artificial intelligence in education by
demonstrating the practical application of AI-driven chatbots in the university admission
process. The primary contributions of this project include:
 Development of an AI-powered Chatbot: The creation of a functional and scalable
Admission Enquiry Chatbot that can assist prospective students with routine
inquiries, improving the efficiency of the admission process.

8
  Integration of NLP and Machine Learning: The integration of Natural Language
Processing (NLP) and machine learning models to enhance the chatbot’s ability to
understand and respond to diverse student queries.
 Evaluation of Effectiveness: Through user testing and feedback, the project
evaluates the chatbot's effectiveness in automating admission inquiries, providing
valuable insights into the benefits and challenges of AI applications in education.
 Improvement of Student Experience: The chatbot contributes to enhancing the
overall student experience by providing 24/7 access to admission-related
information, reducing response times, and ensuring that students receive timely
and accurate responses.

Chapter 2 Literature Review

2.1 chat bot:

A chatbot is a type of software designed to interact with humans in their natural
languages. These interactions can occur in various communication channels such as

9
websites, messaging applications, and social media platforms. Built on artificial
intelligence technology, chatbots can simulate and process human conversation, enabling
them to interact with people in a manner that feels considerably natural.
Chatbots can be programmed to respond in different ways. Some are set up to respond to
specific inputs with pre-defined responses, while others use machine learning algorithms
to adapt their responses to fit the situation. The latter type, also known as AI chatbots, can
learn from past interactions to improve their accuracy over time. Chatbots are widely used
in various industries for tasks such as customer service, ordering products, or providing
information
“Chatbots are computer programs that interact with users using natural languages”

2.2 Learning chatbots

Learning chatbots have the ability to learn from user input. The intelligence of the chatbot
is based on specified rules and by assuming that the user will give a correct answer. There
are two types of learning and both use the same learning rules. The first one will
understand what the user said, will store it for the current session and will reply with that
answer if the user asks about. The second one will store the information in its database
and will reply with that answer for any user. Certain rules can be specified to understand
from previous user input if the possibility of a user saying the truth is high or not and
decide to store the information for future users. Also questions and answers among users
can be saved in the database and compared to identify the possibility of something being
true.

The following conversation gives an example of the learning procedure:

User: How many lecturers are there in the department of computer science?
Chatbot: I am sorry but I don't know the answer. Perhaps you can help me learn
User: There are 28 lecturers at the department
Chatbot: Is this the correct answer?
User: Yes
Chatbot: Thank you for helping
10
User: How many lecturers are there in the department of computer science?
Chatbot: There are 28 lecturers at the department

3.2 Sudanese Universities Using Chat-bots :

A search was conducted on the website of the Sudanese Ministry of Higher Education and
Scientific Research via the link
)https://www.mohe.gov.sd)
to verify which Sudanese universities utilize chat-bot technology in the admission and
registration process, or generally in their interaction with students. Chat-bots are a
modern tool that facilitates communication with applicants and students, providing
immediate responses to their inquiries and contributing to improving the efficiency of
administrative processes.
Upon investigation, it was found that the University of Science and Technology is the only
university that supports chat-bots in its operations. The university uses this technology to
enhance the user experience and provide simplified services to students through its online
platform, making it a leader in this area among Sudanese universities. You can visit the
university’s website through the link (https://ust.edu.sd/ar/)
for more information on how the chat-bot is used in admission and registration
procedures.

2.4 Development of Chatbot-Provided Registration Information Services for Distance

Learning Students:
Focus and Objectives
This study focused on the development and implementation of a chatbot system designed
to provide registration information services for distance learning students. The primary
objectives included:
 Enhancing student satisfaction: Offering a convenient method for accessing
registration details.
 Improving registration processes: Reducing errors and manual workload.
 Leveraging technology: Exploring chatbot applications in education.
 Addressing information gaps: Bridging communication between institutions and
distance learners.
11
Methodology
The research likely employed a design-based approach, which involved:
1. Needs assessment: Identifying specific requirements of distance learners.
2. Chatbot development: Utilizing natural language processing (NLP) for chatbot
design.
3. Implementation: Integrating the chatbot into the registration system.
4. Evaluation: Measuring effectiveness through student satisfaction and operational
efficiency.
Results and Findings
The study found that the chatbot:
 Provided accurate and timely information about registration procedures and
deadlines.
 Improved student satisfaction with its user-friendly interface.
 Significantly reduced administrative workload by automating responses to common
queries.
Chatbot Usage
The chatbot was utilized for:
 Answering FAQs about registration.
 Assisting in troubleshooting technical issues.
 Providing personalized registration details.
Identified Gaps and Proposed Model
The research highlighted a lack of studies on chatbots tailored to distance learners and
proposed a model emphasizing customization, ease of use, system integration, and
iterative improvement.

2. 5 Prototyping a Chatbot for Student Supervision in a Pre-Registration Process:

Focus and Objectives
This study, hosted on ResearchGate, developed a chatbot prototype for guiding students
through pre-registration. The objectives included:
 Streamlining registration processes.
 Enhancing student experience.
12
  Reducing staff workload.
 Exploring chatbot feasibility in academic contexts.
Methodology
Steps likely included:
1. Needs assessment.
2. Chatbot development using machine learning.
3. User testing and feedback analysis.
4. Evaluation of the chatbot's effectiveness.
Expected Outcomes
The study anticipated improvements in:
 Timely and accurate query responses.
 Student satisfaction with chatbot-guided processes.
 Understanding challenges in chatbot deployment and development.
Identified Gap and Proposed Model
A significant gap in chatbot applications for student supervision was identified. The model
proposed integrating NLP and machine learning for enhanced functionality.

2.6 Chatbots in Education and Research: Ethical Implications and Solutions:

Focus and Objectives
This study examined ethical challenges in using chatbots in education and research, such
as:
 Privacy concerns.
 Issues of fairness and accountability.
The study aimed to propose comprehensive solutions to mitigate these challenges.
Methodology
The methodology involved:
1. Literature reviews.
2. Case studies.
3. Expert interviews.

13
 4. Data analysis.
Findings
Key findings included:
 Identification of privacy and transparency concerns.
 Ethical guidelines for responsible chatbot deployment.
 Proposed accountability mechanisms for educational institutions.

2.7 College Enquiry Chatbots Using Iterative Model:

Focus and Objectives
Published in the International Journal of Scientific Engineering and Research, this study
explored chatbot development for college inquiries. Objectives included:
 Enhancing student experience through automation.
 Reducing workload for admissions staff.
 Evaluating iterative development for chatbot improvement.
Methodology
The research employed:
 Iterative development cycles.
 User testing with prospective students.
 Data-driven refinement of chatbot functionality.
Results
 Improved student satisfaction and efficiency in handling inquiries.
 Insights into iterative refinement as an effective development strategy.

Comparative Analysis
Focus Comparison
 Studies 1, 3, and 4 prioritize delivering accurate information to students.
 Study 2 focuses on ethical considerations of chatbot use.
Methodological Comparison

14
  Study 1 and 4 emphasize iterative improvement, while Study 2 employs qualitative
approaches.
 All studies address user-centered design, though Study 2 uniquely highlights ethics.
Impact Comparison
 All studies aim to enhance student satisfaction and institutional efficiency.
 Study 2 uniquely emphasizes ethical responsibility.

Conclusion
These studies demonstrate the transformative potential of chatbots in higher education.
They highlight benefits such as improved efficiency, better student experience, and
operational cost reduction. However, ethical considerations like privacy and fairness
remain critical. Institutions must prioritize user feedback, data security, and iterative
refinement to fully harness the potential of chatbot technology.

Chapter 3: Analysis and design

3.1 Introduction
This chapter provides a comprehensive analysis and design of the Admission Enquiry
Chatbot project. It highlights the system requirements, architecture, and detailed design
considerations to ensure the efficient operation of the chatbot. The project aims to
simplify admission-related inquiries for the College of Computer Science at Sudan
University of Science and Technology by creating an interactive and user-friendly platform.

3.2 Functional Requirements :

User Query Handling:

15
 1. The chatbot should understand and process user queries written in Arabic,
focusing on questions related to admission processes, deadlines, fees, and
required documents.
2. It should be capable of identifying the intent (e.g., "What are the admission
deadlines?") and extracting entities (e.g., dates, document names).

Arabic NLP Support:

1. The chatbot must utilize Rasa’s NLP pipeline to accurately handle Arabic
language inputs, including variations in phrasing and grammar.
2. The model should be trained on Arabic datasets, ensuring it can distinguish
between Modern Standard Arabic (MSA) and common dialects if necessary.

FAQ Database Management:

1. The system should include a backend connected to a MySQL database,
storing frequently asked questions and their corresponding answers.
2. It must allow administrators to update, delete, or add new FAQs dynamically.

Web Integration:
1. The chatbot will be embedded in a web application developed using Flask.
2. It should load as a widget or a dedicated page where users can interact with
it.

Search Functionality:
1. The chatbot should perform database searches to retrieve accurate
responses.
2. Example: A query about application fees will trigger a database query to fetch
the latest fee amount and return it to the user.

Escalation to Human Support:

1. If the chatbot cannot answer a question, it should escalate the query to
human staff by logging the query and the user’s contact information.

Query Logging:
1. Every user query and response must be logged into the database for auditing,
analysis, and improving future interactions.

3.3 Non-Functional Requirements

16
Performance:
1. The chatbot must respond to user queries within 2-3 seconds.
2. Queries to the database should be optimized to minimize delay.

Scalability:
1. The system should support up to 500 concurrent users without performance
degradation.
2. Additional servers or containers (e.g., Docker) can be deployed for load
balancing if needed.

Reliability:
1. Ensure 99.9% uptime for the chatbot service.
2. Implement error-handling mechanisms to gracefully manage unexpected
failures, such as database connection issues.

Security:
1. All data transmitted between users and the chatbot must be encrypted using
HTTPS.
2. Sensitive data, such as user feedback, must be stored securely with
encryption in the database.

Usability:
1. The chatbot interface must be intuitive and accessible to users of all technical
backgrounds.
2. The text input box should guide users by providing examples or hints, such as
"Type your question here, e.g., 'What are the admission requirements?'"

Language Accuracy:
1. Achieve a minimum accuracy rate of 85% in Arabic query understanding
during testing.
2. Regularly retrain the NLP model using new data to improve accuracy and
adapt to user trends.

Cross-Platform Compatibility:
1. The web application should work seamlessly across desktop and mobile
browsers.

17
 Feedback Analysis:
1. Provide analytical tools for university staff to review chatbot performance
metrics, such as the number of queries handled, response accuracy, and
common unresolved questions.

Functional Requirement Example:

A user asks: "‫"ما هي آخر مهلة للتقديم؟‬
 The chatbot detects the intent: Query Admission Deadline.
 It fetches the deadline from the MySQL database and responds:
"2025 ‫ يناير‬1 ‫آخر مهلة للتقديم هي‬."

Non-Functional Requirement Example:

 Performance: The system should respond within 2 seconds, even if 100 users ask
the same question simultaneously.
 Scalability: If user load increases, additional chatbot instances should be deployed
to handle the traffic.

Let me know if you need further clarification or additions!

3.4 System Components:
The Admission Enquiry Chatbot is made up of several key components, each playing an
essential role in providing seamless and efficient user interactions.
1. User Interface (UI):
 The User Interface (UI) is the front-end interface that allows users (prospective
students) to interact with the chatbot.
 This interface is designed to be user-friendly and accessible. It can be implemented
through a web interface or messaging platforms (e.g., WhatsApp, Facebook
Messenger).
 If the interface is web-based, HTML, CSS, and JavaScript will be used to create a
responsive and interactive design that facilitates easy communication with the
chatbot.

2. Chatbot Engine:
 The Chatbot Engine is responsible for processing user input and generating
responses. It uses Rasa Framework for Natural Language Processing (NLP) and
dialogue management.

18
  Rasa is open-source and enables flexible customization, meaning it can be tailored
to understand and respond to specific queries related to university admission
processes.
 The chatbot will handle all incoming inquiries related to the admission process such
as deadlines, eligibility criteria, scholarship details, and course offerings.

3. Database:
 A Database (e.g., MySQL) will store all the critical information related to the
admission process. This includes:
o Application deadlines.

o Eligibility criteria for different courses.

o Available programs and courses.
o Scholarship information.
 The chatbot will query this database to retrieve up-to-date and accurate
information when responding to user queries.
 The database ensures that the information provided is reliable, and can be easily
updated when necessary.

4. Natural Language Processing (NLP):

 The core of the chatbot's ability to understand and process human language is
Natural Language Processing (NLP).
 Rasa NLU (Natural Language Understanding) will be used to train the chatbot to:
o Understand User Intent: The chatbot will identify what the user wants to
know (e.g., "What is the application deadline?" or "What are the available
programs?").
o Extract Entities: The chatbot will identify key pieces of information, such as
the name of the program or the application deadline, from the user's
message.

3.5 Tools and Technologies

 Rasa: For NLP and dialogue management.
 Flask: For backend and API integration.

19
  MySQL: For storing data like deadlines, programs, eligibility, and other admission
information.
 Python: For implementing NLP, Flask backend, and data handling.
 JavaScript, HTML, CSS: For front-end development (if building a web interface).

3.6 Natural Language Processing (NLP) :

Natural Language Processing (NLP) is a subfield of artificial intelligence (AI) that
enables machines to understand, interpret, and generate human language.
1- Applications of NLP:

Chat-bots and Virtual Assistants: Automating customer service and providing

information (e.g., Siri, Alexa).
Sentiment Analysis: Analyzing opinions and emotions in text (e.g., social media
monitoring).
Machine Translation: Translating text between languages (e.g., Google Translate).
Text Summarization: Creating concise summaries of longer texts
Information Retrieval: Enhancing search engines to understand user queries better

2- NLP in Admission Enquiry Chat-bots:

In the context of an admission enquiry chat-bot, NLP helps the chabot in:
Understanding User Intent: The chat-bot uses intent recognition to identify what
the user wants (e.g., checking admission requirements or application status). For
example, when a student asks, "What documents do I need to apply?" the chat-bot
identifies the user's intent to gather information about required documents.
Entity Extraction: NLP allows the chat-bot to extract relevant entities such as dates,
names of programs, or application status (e.g., "application deadline" or
"engineering program"). This information is key to providing personalized and
context-aware responses.
Contextual Conversations: An NLP-based chat-bot keeps track of the conversation's
context to handle multi-step queries. For instance, if a student first asks about
deadlines and later inquires about available programs, the chat-bot can refer back
to earlier responses if needed.

20
 For a Chat-bot, selecting the right NLP framework is crucial for understanding user
queries and providing accurate, real-time responses. Below are some of the best chat-
bot NLP frameworks specifically suited for such use cases:

Library/ Languages Work Features Best for

Framework Supported Environmen
t
NLTK (Natural English, Arabic Python, Pre-trained
Language Toolkit) (via custom Jupyter models,
datasets) dependency
parsing, NER
Rasa Arabic, English Python Open-source, Building custom
NLU, dialogue chat-bots with
management, flexible ML
flexible pipelines models
for different
languages
Microsoft Bot Arabic, English C#, Supports multiple Enterprise chat-
Framework JavaScript, languages, bots
Python integration with
Azure, powerful
NLP models
Dialogflow Arabic, English Web-based, Pre-trained NLP Creating chat-
(Google Cloud models, easy bots with
integration, minimal coding
Google Cloud
support
Botpress JavaScript, Conversational AI Scalable chat-
Node.js platform, bot solutions,
multilingual NLP, enterprise-level
drag-and-drop bots
interface
DeepPavlov English, Arabic Python, Conversational Building
Cloud AI,chat-bot dialogue
frameworks, systems,chat-
context-aware bots
NLP

21
In our Admission Enquiry Chat-bot project, we will use Natural Language Processing (NLP)
technology through the Rasa framework to develop an intelligent chat-bot capable of
understanding and responding to inquiries in Arabic. Rasa is a powerful, open-source
library that allows the development of advanced chat-bots with a high capability to
analyze and understand text, making it ideal for meeting the needs of Arabic-speaking
users.
Why did we choose Rasa?
Arabic Language Support : Rasa enables efficient handling of text written in Arabic
through its built-in Natural Language Processing library. We will use customized models
and configurations for Arabic processing, such as WhitespaceTokenizer and DIETClassifier
, to ensure accurate understanding of user queries.
Flexibility and Customization : Rasa allows us to fully customize the chat-bot, including
the intents and entities that we want the chat-bot to recognize in the students' admission-
related inquiries.
Integration with Other Systems : Rasa can be integrated with internal databases to
provide dynamic information on admission deadlines, tuition fees, and available programs.
Easy integration with messaging platforms (e.g., Whats App, Telegram).

How will we use Rasa in the project?

1. NLP Configuration: We will configure Rasa to process queries written in Arabic by
setting up training models that include expected student questions, such as "What
is the application deadline?" and "What are the tuition fees?".
2. Training the Chat-bot in Arabic : The chat-bot will be trained on a set of Arabic
examples to ensure it can recognize various intents and provide accurate responses
to users.
3. Using Arabic-Specific Libraries : We will utilize libraries like AraBERTor Farasa
Tokenizer to ensure better tokenization of Arabic texts and improved context
understanding, which will enhance the chat-bot response accuracy.
4. Database Integration: The chat-bot will be integrated with our databases to
ensure information is continuously updated, allowing the bot to provide accurate
and up-to-date information on application deadlines and tuition fees.·

3 . 7Use case model

22
The use case diagram describes the functionality of the system as designed from the
requirements and can be found below.

23
Chapter 4: Design

4.1 Introduction:
This chapter provides a detailed explanation of the practical steps taken to develop
the Admission Enquiry Chatbot project. The project aims to deliver an interactive
system based on artificial intelligence techniques to handle students’ inquiries about
admissions and registration. The system was implemented using the Rasa
framework for natural language processing (NLP) and Flask to build a user-friendly
web interface.
The chapter outlines the step-by-step process followed during development, starting
with setting up the programming environment and installing necessary tools,
customizing project files, training the model, integrating the user interface with the
Rasa backend, and thoroughly testing the system to ensure its efficiency.

Implementation Steps
4.2 Setting up the Environment and Installing Tools
1- Creating a virtual environment to manage project dependencies:

python -m venv Adm

2- Activating the virtual environment:

On Windows:
Adm\Scripts\activate
3- Installing the Rasa library:

pip install rasa

4_ Installing Flask for developing the user interface and backend integration:
pip install flask

4.3 Creating a Rasa Project

24
  A Rasa project was initialized using the command:

rasa init
 The project includes essential files such as:
o domain.yml: Defines intents and responses.

o nlu.yml: Provides training data for the model.

o stories.yml: Specifies conversational scenarios.

o config.yml: Configures model algorithms and settings.

3. Customizing Project Files

B. The nlu.yml File:
 Training examples for the intents were included:

intent:‫التحية‬
examples: |
- ‫مرحبا‬
- ‫السالم عليكم‬
- ‫السالم عليكم ورحمة الله‬
- ‫اهال وسهال‬
- ‫هاي‬
A. The domain.yml File:
 Intents such as admission inquiry and program details were added.
 Example responses:

responses:
utter_‫ التحية‬:
- text: "‫"!مرحبا وسهال كيف يمكنني مساعدتك ؟‬
- text: "‫"مرحبا بك في جامعة السودان للعلوم والتكنولوجيا كيف اساعدك؟‬


25
C. The stories.yml File:
 Possible conversation scenarios were defined:

- story: ‫التحية‬
steps:
- intent: ‫التحية‬
- action: utter_‫التحية‬

4. Training the Model

 The model was trained using the following command:

rasa train
 The trained model was saved in the models directory.

4.4 Designing and Integrating the User Interface

with Rasa
A. Designing the User Interface:
 A simple HTML interface was created, including:
o An input field for user queries.

o A section to display chatbot responses.

B. Flask Code for Integration with Rasa:

 The main Flask application code:

from flask import Flask, render_template, request

import requests

app = Flask(__name__)

@app.route('/')
def index():
return render_template('index.html')

@app.route('/chat', methods=['POST'])
def chat():
user_message = request.form['message']
response = requests.post(
'http://localhost:5005/webhooks/rest/webhook',
json={"sender": "user", "message": user_message}
)
return response.json()[0]['text']

if __name__ == '__main__':
app.run(debug=True)

26
27