Title: Estimating the situation of customers' claim for compensation in travel insurances using machine learning

Authors: Yusufcan Yaldiz; Adnan Corum

Addresses: Graduate School, Bahçeşehir University, Ciragan Street, Besiktas, Istanbul, Türkiye ' Department of Industrial Engineering, Bahçeşehir University, Ciragan Street, Besiktas, Istanbul, Türkiye

Abstract: One of the primary concerns for insurance companies is when policyholders file claims after experiencing accidents or other covered events as specified in their policies. Once a claim is initiated, the settlement process usually requires a thorough review, which plays a critical role in influencing customer loyalty and satisfaction. Considering this, the current study seeks to streamline travel insurance claim processes through automation and improve decision-making capabilities. By employing logistic regression and naive Bayes estimation models, using a limited number of variables and an imbalanced dataset, the study evaluates the probability of travel insurance claims. Additionally, principal component analysis is used for reducing dimensions. All analyses are performed using Python and its associated libraries. Ultimately, the results indicate that the implemented models can function as effective decision support tools.

Keywords: travel insurance; insurance claims logistic regression; naive Bayes; principal component analysis; machine learning.

DOI: 10.1504/IJADS.2026.153503

International Journal of Applied Decision Sciences, 2026 Vol.19 No.3, pp.225 - 256

Received: 02 Feb 2024
Accepted: 24 Jul 2024
Published online: 12 May 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article