Medical Insurance

Predicting Medical Insurance Costs

Business Task: Predicting medical insurance costs is crucial for insurance companies to accurately assess risk and set premiums. I developed a model to predict insurance costs based on individual characteristics.

Approach:

• Data Collection: Obtained a dataset containing age, sex, BMI, smoker status, region, and medical insurance costs.

• Data Preprocessing: Handled missing values, encoded categorical variables, and standardized numerical features.

• Exploratory Data Analysis (EDA): Explored the distribution of insurance costs and visualized relationships between features and costs.

• Feature Engineering: Derived new features such as BMI categories and interaction terms.

• Model Selection: Experimented with regression algorithms including linear regression, random forest regression, and gradient boosting.

• Model Training and Evaluation: Trained the model using a dataset split into training and testing sets. Evaluated performance using metrics such as Mean Absolute Error (MAE) and Mean Squared Error (MSE).

Key Findings:

• Identified age, BMI, and smoker status as significant predictors of medical insurance costs.

• Developed actionable insights for insurance companies to refine risk assessment and pricing strategies.

• Achieved a R2 score of 0.86 indicating the model performance is good.

Skills Utilized:

• Programming Languages: Python (NumPy, Pandas, Scikit-learn)

• Tools: Jupyter Notebook, Matplotlib, Seaborn

• Machine Learning Techniques: Regression Analysis

Results: This predictive model provides valuable insights into factors influencing medical insurance costs, enabling insurance companies to make data-driven decisions and optimize pricing models.

GitHub Repository: Link to GitHub Repo