Ngum12/DRIAN---Conflict-Events---Model-Training-and-Evaluation2

Conflict prediction in Africa, this model is just the beep to the big project.., # Conflict Events Classification with Optimization Techniques **Problem Statement:** Africa's crisis management systems face significant challenges due to lack of real-time data integration, insufficient localized context in prediction models, and limited use of multi-modal data sources. These deficiencies lead to delayed and often inaccurate crisis responses. Implementing an AI-driven platform like DRIAN could address these gaps, enhance disaster preparedness, and improve overall crisis response effectiveness across the continent. This project investigates the classification of conflict events using various machine learning approaches. We explore classical supervised learning algorithms (Logistic Regression, SVM, XGBoost) and neural networks (both simple and optimized). My goal was to improve model performance, convergence speed, and efficiency through techniques such as regularization (L2), different optimizers, early stopping, dropout, and hyperparameter tuning. --- ## Dataset & Video Presentation - **Dataset Link**: Click Here to Access the Dataset - **Video Presentation**: Watch the Project Walkthrough In the video, I demonstrate the dataset, code, optimization techniques, and final results in approximately 5 minutes. --- ## Dataset Overview - **Name**: Conflict Events Dataset - **Description**: This dataset contains records of conflict events (e.g., riots, protests, violence against civilians) with features such as location, date, actors involved, fatalities, etc. - **Features**: - **Categorical**: EVENT_TYPE, ACTOR1, COUNTRY, etc. - **Numerical**: YEAR, FATALITIES, LATITUDE, LONGITUDE, etc. - **Target**: `EVENT_TYPE` (multi-class) --- ## Project Structure ``` Project_Name/ ├── notebook.ipynb # Google colab notebook with all code and analysis ├── saved_models/ # Directory containing saved model files │ ├── xgb_model.pkl │ └── optimized_nn_model.h5 └── README.md # This file ``` --- ## Implementation Summary I implemen …