Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2022
|
| Online Access: | https://doi.org/10.5281/zenodo.14769929 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- <p>In this paper, we examine how machine learning techniques can be applied to detect fraud in life insurance and focus on approaches to both supervised learning and anomaly detection. We show a comparison of the differences in performance between models on accuracy, precision, recall, and F1-score of Random Forest, Support Vector Machine, and Isolation Forest. Based on the result, the algorithm of Random Forest performs the best with better accuracy and reached excellent F1 <br>score, thus likely to have great potential in processing large data sets of insurance claims. In addition, applying the anomaly detection technique also seems worthwhile once one has limited data for labeling. There are several obstacles, such as data imbalance, difficulty selecting features, and the need to model interpretability. Furthermore, model updates are required because fraudulent activities change with time. To address these issues, future studies should be conducted to discover new data handling techniques, develop feature engineering processes, and design systems for continuous learning. Machine learning is a game-changer in the life insurance industry because it may allow for better identification of fraudulent claims and increase the trust level of policyholders.</p>