Project Format

1. Introduction (Complete description about the project)

2. Package /Module (Short description of the module)
3. Code
4. Output
5. Requirements
6. Bibliography

Format for Project Abstract Creation

1. Problem Selection –Important issue which you think can be solved using AI by doing survey or interview and
which is aligned to any of the SDGs.
2. Users Affected by the Problem:
3. Empathise – create Empathy Map
4. 6. Define the Problem – Write down the 5W1H questions and write the problem statement
5. Ideate – Brainstorm and write down the ideas
6. Prototype- Create prototype by simple drawing, poster, group role-playing, homemade “gadget, or a 3D
printed product.”
7. Test (optional at this stage)
Project Abstract: Spam Filtering Email System Using AI

 Problem Selection:
Spam emails, also known as junk emails, are unsolicited and often harmful messages that flood our
inboxes daily.
 They can be used for phishing, spreading malware, or simply wasting time.
 Despite advancements in email filtering systems, spam still remains a pervasive issue, causing
inconvenience, loss of productivity, and potential security threats.
 This project seeks to address the problem of spam emails using Artificial Intelligence (AI) to
enhance the accuracy and efficiency of spam detection and filtering, contributing to SDG 9
(Industry, Innovation, and Infrastructure) by fostering innovation in email communication
technologies.

Survey/Interview:

 A survey of email users indicates that 80% of people have received spam emails, with 30% of
them falling victim to phishing attempts.
 Users expressed frustration over the inefficiency of current spam filters and the time spent
manually sorting legitimate emails from spam.

2. Users Affected by the Problem:

 General Email Users: Millions of people around the world use email for communication, with spam
being a constant nuisance.
 Businesses and Professionals: Spam filters can sometimes incorrectly categorize important business
emails as spam, leading to loss of critical communication.
 Organizations: Spam emails can compromise security, leading to data breaches, phishing attacks,
and loss of sensitive information.

3. Empathize - Empathy Map:

To understand the users' pain points, we create an empathy map to visualize their feelings, thoughts, and
needs.

 Says:
o "I spend too much time filtering my inbox."
o "I feel worried when I see a suspicious email."
 Thinks:
o "How do I know if this email is legitimate?"
o "I’m not sure if my spam filter is good enough."
 Does:
o Manually marks emails as spam or moves them to folders.
o Searches for email filters or spam management apps.
 Feels:
o Frustrated and overwhelmed by the constant influx of spam.
o Anxious about missing important emails due to incorrect filtering.
4. Define the Problem – 5W1H Questions & Problem Statement:

 Who: Email users who are affected by the prevalence of spam emails.
 What: Difficulty in effectively filtering spam and distinguishing between legitimate emails and
malicious ones.
 Where: In email inboxes, across various email providers like Gmail, Outlook, Yahoo, etc.
 When: On a daily basis, as spam emails are constantly being sent and received.
 Why: The current spam filters rely on basic algorithms and often fail to detect sophisticated spam or
phishing emails.
 How: Users either manually delete spam emails, risking missing important emails, or continue
dealing with inefficient spam filters.

Problem Statement:
Current spam filtering systems are inefficient, leading to both spam emails slipping through and
legitimate emails being misclassified. This impacts user productivity, creates security risks, and results in
frustration. There is a need for a more advanced AI-based spam filtering system that can accurately
detect spam and prevent it from reaching users' inboxes without incorrectly blocking legitimate emails.
5. Ideate – Brainstorm and Write Down the Ideas:
To solve the spam email issue, several ideas were brainstormed:

 AI-based Email Classifier: Develop a machine learning model that uses natural language
processing (NLP) and pattern recognition to classify emails as spam or legitimate.
 Behavioral Analysis: Implement AI to analyze user behavior and customize spam filtering based on
their email interactions (e.g., frequent contacts, keywords, or domains).
 Phishing Detection System: Introduce AI algorithms specifically trained to detect phishing emails,
such as those containing malicious links or requesting sensitive information.
 User Feedback Loop: Allow users to "train" the AI by marking emails as spam or not spam,
enabling the system to continuously improve its accuracy over time.
 Hybrid Approach: Combine traditional rule-based filtering with AI learning to create a robust,
adaptable filtering system.

6. Prototype – Create Prototype by Simple Drawing, Poster, Group Role-playing, Homemade Gadget,
or 3D Printed Product:

 Poster: Create a visual representation of the AI-powered spam filter, showing how the system works.
The poster will feature a flowchart showing:
1. Incoming Email: Email arrives in the inbox.
2. AI Spam Filter: AI analyzes the email’s content, sender, subject line, and other attributes.
3. Decision: Email is either marked as spam or moved to the inbox.
4. User Interaction: If the email is marked as spam by the user, the AI improves based on
feedback.
 Group Role-Playing: A skit can demonstrate how users interact with the spam filter system. One
group member could be an email user, and another could play the role of the AI spam filter. They
could show how an email is processed, how spam is filtered, and how the system improves over time
based on feedback.
 Homemade "Gadget": Build a basic email filtering system using free AI tools (such as Google's
TensorFlow) to simulate how AI-based filters would work on a small scale, helping to understand the
logic behind AI spam filtering.

7. Test:
Once the prototype is developed, the next step is to test it:

 Internal Testing: Test the AI system on a small dataset of real and spam emails to evaluate its
accuracy. Measure how many legitimate emails are misclassified and how well the AI detects new
types of spam.
 User Testing: Gather feedback from a sample group of users who would use the spam filter in their
daily email routines. Measure their satisfaction with the accuracy and efficiency of the filter, and
make improvements based on their feedback.
 Performance Evaluation: Track the reduction in spam emails and the improvement in user
productivity and satisfaction after using the AI-powered spam filter.