miirshe/Terrorism-Text-Classification

A machine learning and deep learning project to detect and classify Somali-language text as **terrorist** or **non-terrorist**. # 🇸🇴 Somali Terrorism Text Classification A machine learning and deep learning project to detect and classify Somali-language text as **terrorist** or **non-terrorist**. The project explores a range of traditional and neural models to assist in automated threat detection in Somali content. --- ## 🧠 Models Used This project evaluates the following classification models: - ✅ LightGBM with Bag-of-Words (BoW) - ✅ MLP Classifier with TF-IDF - ✅ CNN without Word2Vec - ✅ RNN (LSTM) without Word2Vec - ✅ CNN with Word2Vec Embeddings --- ## 📈 Model Accuracy & Performance ### 🔹 LightGBM (BoW) --- ### 🔹 MLP (TF-IDF) --- ### 🔹 CNN (No Embedding) --- ### 🔹 RNN (LSTM) --- ### 🔹 CNN + Word2Vec --- ## ✍️ Authors - **Miirshe** — github.com Built by Somali engineers and students committed to advancing NLP and public safety with AI.