Institutional Repository of Vilnius University
Klasifikavimo modelių taikymas draudimo sukčiavimo aptikimui esant klasių disbalansui /
Abstract
dc:descriptionThis thesis investigates the detection of fraudulent car insurance claims using a publicly available dataset. The aim of the study was to apply and compare different classification models for fraud detection. The main tasks included analyzing scientific literature on fraud detection methods, applying classification models, optimizing model hyperparameters using Optuna, and evaluating feature importance using SHAP values. After conducting data analysis and correcting data inconsistencies, the dataset was prepared for modeling. Three classification models were applied: logistic regression, random forest classifier, and XGBoost. To address class imbalance, the decision threshold was adjusted. The results showed that all models were capable of performing classification, however they struggled to identify stronger fraud patterns and to clearly distinguish between fraudulent and nonfraudulent claims.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Samušytė, Rugilė,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD308334954&prefLang=en_US