SIRGIANE/nlp_project

Modeling Emotions & Their Intensities ​ on Low-Resource Dialectal ​ Arabic Texts​ # Emotion Intensity Modeling in Low-Resource Arabic Dialects *A Comparative Study of Classical ML and Deep Learning Models* ## Overview This repository contains code and resources for emotion classification and intensity prediction in Moroccan Darija and Algerian Arabic texts. The project compares: - **Classical ML models** (Naive Bayes, SVM, Random Forest) - **Deep Learning models** (LSTM, BERT variants: AraBERT, mBERT, XLM-R) ## Key Features - Custom preprocessing pipeline for Arabic dialects (`DarijaPreprocessor.py`) - Multi-label emotion classification (anger, disgust, fear, joy, sadness, surprise) - Intensity prediction (low/medium/high) - Fine-tuned transformer models (AraBERT, mBERT, XLM-R) ## Repository Structure: ## Installation ```bash git clone github.com cd nlp_project