Spam Email Detection Analysis

Project Overview

This project implements a machine learning pipeline to detect spam emails using the spam_email_dataset.csv. It includes data cleaning, feature engineering (TF-IDF for text and Min-Max scaling for numeric data), and model comparison.

Dataset

The dataset contains email metadata and content, labeled as either 'spam' (1) or 'ham' (0).

Models and Performance

We compared two primary classification algorithms:

1. K-Nearest Neighbors (KNN):

   • Hyperparameters were tuned using cross-validation.
   • Optimized K: 2
   • Achieved approximately 86% accuracy.

2. Multinomial Naive Bayes:

   • This model performed exceptionally well on the processed text data, achieving near 100% accuracy in this specific evaluation.

How to Use

1. Ensure spam_email_dataset.csv is in the root directory.
2. Run the notebook cells sequentially to reproduce the analysis and performance reports.

Requirements

• Python 3.x
• Pandas
• NumPy
• Scikit-learn
• Matplotlib/Seaborn