Comparison of IQR and Isolation Forest Methods for Outlier Detection in Insurance Claim Data at PT Askrindo Semarang Branch
DOI:
https://doi.org/10.26623/transformatika.v23i2.13726Abstract
This study aims to compare two commonly used approaches for outlier detection, namely the statistical Interquartile Range (IQR) method and the machine learning–based Isolation Forest method, using a case study of insurance claim data from PT Askrindo Semarang Branch. The analyzed data include claim amounts and claim settlement duration. The univariate IQR method identified 9 claims classified as outliers, consisting of 5 claim amount outliers, 3 settlement duration outliers, and 1 combined outlier. Meanwhile, the multivariate Isolation Forest method with a contamination rate of 0.05 detected 6 anomalous claims, including extremely long settlement durations and unusual combinations of variable patterns. The results indicate that the IQR method is effective for rapid detection of extreme values, while Isolation Forest is more suitable for identifying complex anomalies across multiple variables. Therefore, the choice between IQR and Isolation Forest methods should be adjusted according to the analysis objectives and data characteristics.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 jovita wayan ramadhani, susanto
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
Transformatika is licensed under a Creative Commons Attribution 4.0 International License.
