Nardosk30/classify-Amharic-News

it is used to classify Amharic news based on their category # 📘 Amharic News Classification System ## 📰 Project Title **Amharic News Classification with Visualization and Machine Learning** --- ## 🎯 Objective This project aims to develop a complete **news categorization system** for Amharic-language news articles using **Natural Language Processing (NLP)** and **Machine Learning (ML)** techniques. It supports end-to-end processing from data cleaning and visualization to model training and evaluation, with proper support for **Amharic fonts and visualizations**. --- ## 📁 Dataset - **Source**: [Local Amharic News Dataset] - **Format**: CSV - **Columns**: - `headline`: Title of the news article - `category`: Numeric label for the news category (e.g., 0000, 0001, etc.) - `date`: Publication date - `views`: Number of views (optional) - `article`: Full article text - `link`: Source URL (optional) --- ## 🧰 Technologies Used | Category Tools | Language Python 3.10+ | NLP NLTK | ML Models Multinomial, NaiveBayes | Vectorizer TF-IDF | Data Handling pandas, numpy | Visualization matplotlib, seaborn | Font Handling Abyssinica SIL / Ethiopic Fonts | Model Persistence joblib | Progress UI tqdm | Jupyter Support IPython display --- Pipeline Overview ### 🔠 Step 1: Font Configuration Detect and register **Amharic fonts** or fallback to system fonts. ### 📥 Step 2: Dataset Loading and Cleaning Load dataset, handle missing values, and map categories. ### 🧼 Step 3: Text Preprocessing Clean text using Unicode normalization and Amharic character filtering. ### 📊 Step 4: Exploratory Data Analysis Visualize category distribution and text lengths using seaborn and Amharic fonts. ### 🧪 Step 5: Feature Engineering Extract features using TF-IDF vectorization with n-gram …