Hybrid Framework for Email Threat Classification and Detection

Project Title

Hybrid Framework for Email Threat Classification and Detection

Project Area of Specialization Cyber SecurityProject Summary

In this age, the most professional and powerful way of communication is email. The institutions like universities, companies and other official sectors transfer and receive tremendous emails on daily basis. With the immense use the emails have been the source to crack the user system for the hackers. Most emails we receive include malicious part harmful for our system. To overcome this issue, a number of email filtering systems are available. But they have certain limitations like they only work on rule based systems and only focus on one part of email. Their prediction lacks due to limitation of less component analysis. We are addressing this issue by foucusing on three main modules of email for Detection. These include text, URLs and attachments. For better prediction and analysis, we are using Machine learning, Deep learning and

Natural Language processing techniques. This would give the more accurate prediction than previous email detectors. Along with this, our framework would also be able to classify the malicious part. It means the model would detect which part of email is a harmful for system and then to what extent this would be harmful.

Our FYP group consists of 2 members so both of us are equally involved in each step from coding to research and from documentation to presentation.

Project Objectives

Our main purpose is to make a framework for detecting any email threat and classifying it to
reduce any harm that can be caused to users systems. One purpose is to make our system user
friendly by combating with proper user interface. Following are our goals for completing our
project:
1.To detect harmful messages in email and scanning for any unwanted and unsolicited
mails.
2.Preventing malware containing files from reaching users system.
3.Checking for phishing sites links and preventing user from clicking on them by alerting
him.
4.Preventing a user from receiving messages from a person who has any past crime or
hacking record.
5.If any section of email, even if its the subject of the email, is harmful or if it contains bad
words then, alerting user about it.
6.Classifying the threat in the email and telling that where does the problem exist.

Project Implementation Method

Methodology
We’ll use research by different experts in combat with our own research. Content will be
analysed using programming language(s). Every section of email will be detected for any threats
using many techniques for analyzing the threat and then classifying the kind of threat in that
email. Learning from multiple techniques from previous research will also be implemented on
each section of mail. Accuracy and precision will be checked for testing and analyzing the
results. Data that would need visualisation will also be visualized in order to create a better
understanding of results.

1. Design
1.1.  Preparing Data
Data will be prepared by getting from internet and on our own. It will be analyzed and visualized
if needed. Also if the data will need any splitting or preprocessing then it will be applied.
1.2. Data Preprocessing
Data cleaning, data reduction, data transformation and data integration may be done to extract
the information that should be included in final data that would be processed.
2. Implementation Steps
2.1. Applying Different Techniques
Different techniques will be applied for getting the desired framework.
2.2. Analyzing Results
The final report will be analyzed after visualizing the data and results that need visualization.

Benefits of the Project

Socio-Economic benefits include:
1.Users will be saved from any kind of scam or malicious activity.
2.Thehigherorganizationsthatarethebiggesttargetofhackerswillbesaved from
information leakage.
3.The additional feature that involves alerting the user about the threat would be significant for
busy persons.
4.The accuracy in filtering spam emails would be increased due to the use of machine learning
algorithms.

Technical Details of Final Deliverable

SAS (Software as a service)

Final Deliverable of the Project Software SystemCore Industry SecurityOther Industries IT Core Technology OthersOther Technologies Artificial Intelligence(AI)Sustainable Development Goals Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	78000
Computer System	Equipment	1	60000	60000
Hard Drive	Equipment	1	8000	8000
Dataset from company	Miscellaneous	1	10000	10000