enron-sender-detection

A simple machine learning approach to detect the sender based on the mail body of the famous Enron Datasset

prerequisites

• python
• anaconda
• scikit learn
• other dependencies

How to run

1. clone this repository - git clone https://github.com/nowshad-sust/enron-sender-detection.git
2. download enron dataset from here - https://www.cs.cmu.edu/~./enron/enron_mail_20150507.tgz
3. now extract this dataset(maildir) to the project(clonned) folder
4. create a folder named remail in the project directory
5. open a terminal or cmd in the project directory
6. run the copy_sent_mails.py script by the command - python copy_sent_mails.py This should make a directory named remail in the project folder and copy all the sent mails from the original dataset directory.
7. Now, run the naive_bayes_pipeline.py by the command - python naive_bayes_pipeline.py This should give you a number which refers to the validation sucess rate.

Latest Statistics (accuracy)

• Naive Bayes classifier ~ 0.46
• SVM ~ 0.79
• SVM with grid search ~ 0.85