Development of a Risk-Scoring System to Predict Car Insurance Fraud
Date
2-13-2024
Faculty Mentor
Julie Staples, Finance, Economics & Accounting; Keith Lowe, Finance, Economics & Accounting
Loading...
Files
Submission Type
Conference Proceeding
Location
4:15-4:25pm | Houston Cole Library, 11th Floor
Description
Car insurance fraud constitutes a substantial threat to insurance companies, resulting in significant financial losses. Proactive fraud prevention plays a crucial role in mitigating financial losses. Machine learning approaches are frequently used to predict car insurance frauds. However, the machine learning approach is often viewed as nebulous, posing challenges in interpretation, particularly in the realms of accounting and finance. In contrast, point-based scoring systems (like credit scores) are becoming more widespread in sectors such as healthcare and finance due to its easy interpretation. The objective of this study is to develop a simple risk score system for car insurance fraud that is easily understood. The scoring deviation framework used in disease prediction is adapted for car insurance fraud prediction. The scoring system generates a table containing points for each risk factor of insurance fraud. To evaluate the performance of the proposal, real-life data from a Kaggle dataset containing 15,420 records was utilized to train and test the model. This method not only examines scores but is also useful for insurance professionals to evaluate validity.
Keywords
student research, finance
Rights
This content is the property of Jacksonville State University and is intended for non-commercial use. Video and images may be copied for personal use, research, teaching or any "fair use" as defined by copyright law. Users are asked to acknowledge Jacksonville State University. For more information, please contact digitalcommons@jsu.edu.
Disciplines
Finance and Financial Management
Recommended Citation
Gao, Qian, "Development of a Risk-Scoring System to Predict Car Insurance Fraud" (2024). JSU Student Symposium 2024. 40.
https://digitalcommons.jsu.edu/ce_jsustudentsymp_2024/40