A framework for detecting Spoofed Emails by using python
With the internet becoming an ever evolving technology, ways to protect ourselves from malicious attacks have become ever so important.
Everyday these attacks become more and more intelligent in trying to trick not only the user but also current spam protection systems.
This illustrates that most current implementations lack in detecting the hijacking of already “trusted” email addresses.
Having a system capable of analysing and detecting such malicious attacks would be of great use.
This project aims to create such a system and explore the multiple ways in which Artificial Intelligence and strict filtering rules can leverage advancements in the field of Cyber security and Datascience .
Sea stands for Spam . Email - Analysis. It can be used to train and test large amounts of data and also classify single Subject HEadings and Email addresses. Compared to more traditional methods , SEA justifies its classification using both the Subject heading , Email address while also passing through custom filters that further filter out smaller details not addressed by the classification models.
Using our custom filtering techniques the overall precision of the software increases drastically.
Autamated installation can be done using the start.sh file.
Due to githubs file size limitation this cannot be included in the repository , so download , unzip and move the
files into the Prototype/loader folder
wget http://nlp.stanford.edu/data/glove.6B.zipunzip glove.6b.zip
sudo pip3 install inquirer tqdm colorama nltk pandas autocorrect pympler keras tensorflow keras_metrics sklearn ann_visualizer pyfiglet textblob
python3 -m textblob.download_corpora
cd /Prototype/python3 -W ignore start.py
Place the files in the dataset folder and change the path in the respective files you wish to train with.
There is no guarantie that your dataset will be read by the program, as some datasets use different delimeters for spacing out the data.