🚀 Building a Phishing Email Detector using Machine Learning

📌 Introduction

Cybersecurity is one of the most critical concerns in the digital era, and phishing attacks remain a major threat. As a final-year Cybersecurity student, I wanted to apply my skills and create something practical and relevant. That’s how this Phishing Email Detector project was born.

🔍 Problem Statement

Phishing emails trick users into revealing sensitive information by mimicking trusted sources. Manual detection is time-consuming and error-prone. My goal was to automate this process using machine learning.

🛠️ Tech Stack

• Python

• Flask (for web app)

• Scikit-learn (ML modeling)

• Pickle (model storage)

• HTML/CSS (Frontend)

• GitHub (Version control)

📂 Dataset

The dataset contains labeled emails:

• 0 → Legitimate

• 1 → Phishing

Each email text was preprocessed and vectorized using TF-IDF to extract meaningful features.

🤖 Model Training

• Model: Logistic Regression

• Achieved 100% Accuracy on test data (small sample for prototype)

• Tools: train_model.py script handled preprocessing, training, evaluation, and model saving.

💾 Output

Model and vectorizer were saved as:

model/phishing_model.pkl  
model/tfidf_vectorizer.pkl

🌐 Web Interface

Built a simple yet functional web interface using Flask and HTML.
Users can:

1. Paste an email message

2. Click "Check"

3. Get immediate result: Phishing or Not

📦 GitHub Repository

Here’s the full source code:
👉 github.com/itz-krithick-raj/phishing-detector

🔐 Future Improvements

• Use deep learning with LSTM models

• Expand dataset for higher reliability

• Add email header analysis

• Implement real-time scanning from inboxes

🧠 What I Learned

• End-to-end ML deployment

• Flask web development

• Git/GitHub workflows

• Debugging and problem-solving during deployment