Modeling Car Insurances

Overview

This project aims to develop a predictive model to estimate the likelihood of a customer making a claim on their car insurance during the policy period. Commissioned by a fictional car insurance company, the project seeks to optimize pricing strategies and enhance risk assessment capabilities, crucial for maintaining a competitive edge in the large car insurance market.

Table of Contents

• Data Contents
• Methodology
• Results
• Visualizations
• Conclusion

Data Contents

The dataset, car_insurance.csv, includes various customer attributes:

Column	Description
id	Unique client identifier
age	Client's age: 0: 16-25 1: 26-39 2: 40-64 3: 65+
gender	Client's gender: 0: Female 1: Male
driving_experience	Years the client has been driving: 0: 0-9 1: 10-19 2: 20-29 3: 30+
education	Client's level of education: 0: No education 1: High school 2: University
income	Client's income level: 0: Poverty 1: Working class 2: Middle class 3: Upper class
credit_score	Client's credit score (between zero and one)
vehicle_ownership	Client's vehicle ownership status: 0: Does not own their vehilce (paying off finance) 1: Owns their vehicle
vehcile_year	Year of vehicle registration: 0: Before 2015 1: 2015 or later
married	Client's marital status: 0: Not married 1: Married
children	Client's number of children
postal_code	Client's postal code
annual_mileage	Number of miles driven by the client each year
vehicle_type	Type of car: 0: Sedan 1: Sports car
speeding_violations	Total number of speeding violations received by the client
duis	Number of times the client has been caught driving under the influence of alcohol
past_accidents	Total number of previous accidents the client has been involved in
outcome	Whether the client made a claim on their car insurance (response variable): 0: No claim 1: Made a claim

Methodology

1. Data Cleaning and Transformation: Handled missing values and transformed categorical variables into numerical representations.

2. Exploratory Data Analysis (EDA): Analyzed correlations and visualized data distributions to understand feature relationships.

3. Feature Selection: Applied logistic regression to assess each feature's predictive power.

4. Modeling:

   • Logistic Regression: Tuned using GridSearchCV, achieving an accuracy of 79%.
   • Random Forest: Tuned with parameters max_depth: None, min_samples_split: 2, n_estimators: 50, achieving an accuracy of 82%.

5. Evaluation: Used confusion matrices and classification reports to evaluate model performance.

Results

• Best Feature: driving_experience was identified as the most predictive feature with an accuracy of 77.71%.
• Model Performance:
  • Logistic Regression: Precision of 0.92 for "No Claim" and 0.63 for "Claim".
  • Random Forest: Precision of 0.85 for "No Claim" and 0.70 for "Claim".

Visualizations

Confusion matrices and correlation heatmaps are included to provide visual insights into model performance and feature relationships.

Conclusion

The project successfully identified key predictive features and developed models that enhance decision-making processes. The Random Forest model, with its higher accuracy, is recommended for deployment, providing a balance between simplicity and performance. This approach allows the company to start with a straightforward model in production, minimizing the need for complex infrastructure and expertise.

Getting Started

To run this project, ensure you have the following libraries installed:

• pandas
• numpy
• matplotlib
• seaborn
• scikit-learn

Clone the repository and run the Jupyter Notebook to explore the analysis and predictions.

License

This project is licensed under the MIT License - see the LICENSE file for details.