EstifanosTeklay/AlphaCareInsurance

The project is dedicated to advancing risk and predictive analytics within car insurance planning and marketing in South Africa. # AlphaCareInsurance # Insurance Claims & Premium Optimization using Machine Learning This project applies statistical modeling and machine learning techniques to analyze insurance claims and optimize premium pricing for AlphaCare Insurance Solutions. The goal is to identify key risk drivers and develop predictive models that support data-driven underwriting and pricing decisions. ## 📌 Business Problem Accurate premium pricing and claims prediction are critical for insurance profitability and customer retention. Traditional pricing models often fail to capture complex non-linear relationships between policy, vehicle, customer, and location attributes. This project addresses: - Claims prediction at policy and location level - Premium optimization using machine learning - Identification of key risk and pricing drivers ## 📊 Dataset Overview The dataset contains over **618,000 insurance policy records**, including: - Customer demographics and financial attributes - Vehicle characteristics and age - Policy and coverage details - Geographic risk indicators - Historical premium and claims information Target variables: - `TotalClaims` - `TotalPremium` ## 🔧 Methodology 1. **Data Cleaning & Preparation** - Missing value treatment - Feature engineering (vehicle age, power ratios, loss metrics) - Categorical encoding - Train-test split (80/20) 2. **Modeling Techniques** - Linear Regression (baseline) - Random Forest Regression - XGBoost Regression 3. **Evaluation Metrics** - RMSE - MAE - R² Score 4. **Feature Importance Analysis** - Tree-based feature importance - Business interpretation of drivers ## 🤖 Models Implemented | Model | Purpose | |-------------------|--------| | Linear Regression | Baseline & interpretability | | Random Forest | Non-linear modeling & feature importance | | XGBoost | High-performance predictive modeling | XGBoost delivered the best performance for both claims and premium prediction. ## 📈 Key Findings - Tree-b …