Elizabeth-Mwania/NollySenti_en_yo_ha

NollySenti_en_yo_ha is a machine learning project dedicated to analyzing and classifying sentiment in Nollywood content. This project leverages multilingual NLP techniques to understand audience sentiment across three major Nigerian languages: English, Yoruba, and Hausa. # NollySenti: Multilingual Sentiment Analysis for Nollywood Films A comprehensive sentiment analysis project exploring emotional patterns in Nigerian cinema across multiple languages (English, Yoruba, and Hausa). ## 📋 Overview **NollySenti_en_yo_ha** is a machine learning project dedicated to analyzing and classifying sentiment in Nollywood (Nigerian film industry) content. This project leverages multilingual NLP techniques to understand audience sentiment across three major Nigerian languages: English, Yoruba, and Hausa. ### Key Features - **Multilingual Support**: Sentiment analysis in English, Yoruba, and Hausa - **Nollywood Focus**: Specifically trained on Nigerian film industry content - **Comprehensive Dataset**: Includes training, development, and test datasets - **GPU-Optimized**: Runs on Google Colab with T4 GPU acceleration - **Deep Learning Approach**: Implements transformer-based models for accurate sentiment classification ## 📊 Project Structure ``` NollySenti_en_yo_ha/ ├── NollySenti_en_yo_ha.ipynb # Main analysis notebook └── README.md # Project documentation ``` ## 🗂️ Datasets The project utilizes TSV-formatted datasets for training and evaluation: - **train.tsv**: Training dataset for model development - **dev.tsv**: Development/validation dataset for hyperparameter tuning - **test.tsv**: Test dataset for final model evaluation Each dataset contains labeled examples of film-related text with corresponding sentiment labels. ### Dataset Source - **Primary Dataset**: Davlan/nollysenti - A comprehensive multilingual Nollywood sentiment analysis dataset available on Hugging Face ## 🛠️ Technical Stack - **Language**: Python 3 - **Deep Learning Framework**: PyTorch / Transformers - **Computation**: GPU-accelerated (T4 GPU on Google Colab) - **Notebooks**: Jupyter Notebook (IPython) - **Data Format**: TSV (Tab-Separated Values) ### Pre-trained Models - **Model**: Davlan/afro-xlmr-base - XLM-RoBERTa model fine-tuned f …