Spam Email Classification

Introduction

In today's digital world, spam emails are a constant nuisance. They clog inboxes, waste time, and can even pose security threats containing phishing attempts or malware. This document details the development of a machine learning model designed to tackle this problem by automatically classifying incoming emails as spam or legitimate (ham).

Project Description

This is a Machine learning model to classify emails as spam or non-spam, leveraging Python and popular libraries. The project involved preprocessing email text data, extracting relevant features, and training classification algorithms to achieve high accuracy in spam detection.

• Data Overview

1. The dataset has five columns:
2. Index: A unique identifier for each email.
3. #Sent Emails: How many times the email has been sent.
4. Label: Whether the email is spam or legitimate.
5. Text: The content of the email.
6. Binary Label: A 0 or 1 representing spam or not.

Key Achievements

• Data Preprocessing: Cleaned and prepared raw email data, including tokenization, stop-word removal, and feature extraction using techniques like TF-IDF (Term Frequency-Inverse Document Frequency).

• Model Training: Implemented and evaluated various classification algorithms such as Naive Bayes, Support Vector Machines (SVM), and Logistic Regression to determine the most effective model for spam detection.

• Data Exploring and Visualizations: Provided charts to better understand the data:

  1. Histogram

2. Pie Chart

3. Word cloud for Spam emails

4. Word cloud for ham emails

5. Top 30 words in Spam Emails

6. Top 30 words in ham emails

Model Comparison:

Tested several algorithms to see which works best:

Naïve Bayes

• Training Accuracy: 94.00%

• Testing Accuracy: 93.33%

• Confusion Matrix:

Logistic Regression

• Training Accuracy: 97.00%

• Testing Accuracy: 96.91%

• Confusion Matrix:

SVM

• Training Accuracy: 97.34%

• Testing Accuracy: 97.29%

• Confusion Matrix:

Decision Tree

• Training Accuracy: 99.85%

• Testing Accuracy: 95.60%

• Confusion Matrix:

KNN

• Training Accuracy: 95.14%

• Testing Accuracy: 90.62%

• Confusion Matrix:

• Deployment: Created a simple user interface to input emails and classify them in real-time

Contributors

Zeina Wady

Sara Darwish

Ruba AbdELSalam

Basmala Ayman

Sara Habib

Bassant Ahmed