Eldiyanaa/EthioMart-Amharic-Named-Entity-Recognition

# EthioMart-Amharic-Named-Entity-Recognition ## Overview EthioMart aims to become the primary hub for all Telegram-based e-commerce activities in Ethiopia. With the rising popularity of Telegram as a platform for business transactions, this project seeks to consolidate multiple independent e-commerce channels into a single, centralized platform. The goal is to provide a seamless experience for customers to explore and interact with various vendors in one place. ## Business Need With the decentralization of e-commerce channels, both vendors and customers face challenges in product discovery, order placement, and communication. EthioMart plans to develop a system that aggregates real-time data from various Telegram channels, focusing on fine-tuning a Large Language Model (LLM) for Amharic Named Entity Recognition (NER). This will enable the extraction of key business entities such as product names, prices, and locations from shared content. ## Key Objectives - Real-time data extraction from Telegram channels. - Fine-tune LLM to extract entities like product names, prices, and locations. ## Data Sources - **Messages and data** from Ethiopian-based e-commerce Telegram channels. - Sample data collected from Shageronlinestore link. - Amharic news labeled NER dataset. ## Knowledge and Skills Required 1. **Text Processing**: Handling Amharic text, tokenization, and preprocessing techniques. 2. **LLM Fine-tuning**: Adapting large language models for Amharic NER tasks. 3. **Model Comparison & Selection**: Evaluating performance using metrics like F1-score, precision, and recall. 4. **Model Interpretability**: Using tools like SHAP and LIME to explain model predictions. ## Outcomes - A working pipeline for entity extraction from Amharic Telegram messages. - A performance analysis of different models and their interpretability. - Insights into how the extracted entities can be used for business intelligence in e-commerce contexts. ## Tasks Overview ### Task 1: Dat …