🔗 Spam Detection from URLs using NLP & SVM

This repository presents a Natural Language Processing (NLP) project aimed at detecting spam URLs using machine learning. The project performs text cleaning, feature extraction, and classification using an SVM model with grid search tuning.

---

📁 Project Structure

📦 URL-Spam-Detection/
├── explore.ipynb         # Jupyter notebook with full analysis
├── README.md                 # Project description and usage

---

📊 Dataset

• 📂 Source: 4Geeks Academy GitHub - url_spam.csv
• 💡 Contents:
  • URL text
  • Spam labels (binary classification)
• 🔍 Initial cleanup removed over 600 duplicate entries

---

🧹 NLP Preprocessing

• ✅ Lowercasing, stopwords removal (nltk)
• 🔤 Lemmatization with WordNetLemmatizer
• 🔧 Regex cleanup
• ☁️ WordCloud for spam vs. non-spam terms

---

🧠 Model & Evaluation

• ✨ Model: Support Vector Machine (SVM)
• 🔍 Tuning: GridSearchCV with RepeatedStratifiedKFold
• 📈 Metrics: Classification report including Precision, Recall, F1-score

---

🚀 How to Run

1. Clone the repository

git clone https://github.com/jilemp/URL-Spam-Detection.git
cd URL-Spam-Detection

2. Install dependencies

pip install -r requirements.txt

3. Launch the notebook

jupyter notebook "explore.ipynb"

---

📚 Libraries Used

• pandas, numpy
• nltk, regex
• matplotlib, wordcloud
• sklearn

---

📬 Contact

Project developed as part of 4Geeks Academy — NLP Curriculum. Feel free to open an issue or fork the repo for improvements!