Performance of Prediction Interval Estimators based on Random Forest Models with Correlated Predictors
DOI:
https://doi.org/10.25077/jmua.14.4.320-332.2025Keywords:
Prediction interval, Random Forest, Quantile Regression Forest, Out of Bag Interval Prediction, Split Conformal Prediction, MulticollinearityAbstract
Uncertainty in prediction results is a crucial aspect that needs to be taken into account in regression modeling, especially when there is a high correlation between explanatory variables. This study aims to evaluate the performance of three prediction interval formation approaches, namely Out-of-Bag Prediction Interval (OOB-PI), Quan tile Regression Forest (QRF), and Split Conformal Prediction (SC), in Random Forest modeling. The evaluation was conducted through a simulation study with a variety of data structures, including the level of correlation between variables, the shape of the mean function, and the type of error distribution. Further validation was conducted using data from the National Socio-Economic Survey (SUSENAS) of West Java Province in 2023. The results show that increasing the correlation between explanatory variables can improve the efficiency and accuracy of prediction interval estimation. Overall, OOB-PI showed the most balanced performance compared to the other two methods, with a prediction coverage rate close to 90% and a narrower interval width than QRF and SC. This finding indicates that OOB-PI is an adaptive and efficient approach for various data structures, including socioeconomic data with highly correlated predictors.
Downloads
References
[1] James, G., Witten, D., Hastie, T., Tibshirani, R., 2021, An Introduction to Statistical Learning, Springer Texts in Statistics, Springer, New York.
[2] Gregorich, M., Strohmaier, S., Dunkler, D., Heinze, G., 2021, Regression with Highly Correlated Predictors: Variable Omission is Not the Solution, International Journal of Environmental Research and Public Health, Vol. 18(8): 1–15.
[3] Bickel, P., Diggle, P., Fienberg, S., Gather, U., Olkin, I., Zeger, S., 2017, Springer Series in Statistics, Springer, Berlin.
[4] Fife, D. A., D’Onofrio, J., 2023, Common, Uncommon, and Novel Applications of Random Forest in Psychological Research, Behavior Research Methods, Vol. 55(5): 2447–2466.
[5] Johnson, R. A., 2024, Quantile-Forest: A Python Package for Quantile Regression Forests, Journal of Open Source Software, Vol. 9(93): 5976.
[6] Oliveira, R. I., Orenstein, P., Ramos, T., Romano, J. V., 2024, Split Conformal Prediction and Non-Exchangeable Data, Proceedings of Machine Learning Research, Vol. 25: 1–15.
[7] Zhang, H., Zimmerman, J., Nettleton, D., Nordman, D. J., 2020, Random Forest Prediction Intervals, Statistical Modelling, Vol. 20(5): 481–500.
[8] Harum, N. S., Aini, M., Risxi, M. A., Kartiasih, F., 2023, Pengaruh Sosial Ekonomi dan Kesehatan terhadap Pengeluaran Konsumsi Pangan Rumah Tangga Provinsi Jawa Tengah Tahun 2020, Seminar Nasional Official Statistics, Vol. 1: 899–908.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Departemen Matematika dan Sains Data FMIPA UNAND
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All articles published in Jurnal Matematika UNAND (JMUA) are open access and licensed under the Creative Commons Attribution-ShareAlike 4.0 International License (CC BY-SA 4.0). This license allows published content to be freely accessed, shared, redistributed, and adapted for any purpose, including commercial use, provided that the applicable license conditions are followed.
Copyright Holder
The copyright of articles published in Jurnal Matematika UNAND is held by the Departemen Matematika dan Sains Data, Fakultas Matematika dan Ilmu Pengetahuan Alam (FMIPA), Universitas Andalas (UNAND), in accordance with the journal's stated copyright policy.
Author Rights
Under the CC BY-SA 4.0 license, authors may share, distribute, and reuse their published articles, subject to the applicable license conditions. These rights include:
-
Self-Archiving: Authors may deposit their published articles in institutional repositories, personal websites, or academic platforms.
-
Reuse: Authors may reuse their published work in subsequent publications, research, teaching materials, or other lawful purposes, provided that appropriate attribution is maintained.
-
Distribution: Authors may share and distribute their published articles without requiring additional permission under the terms of the license.
Creative Commons License (CC BY-SA 4.0)
Under the CC BY-SA 4.0 license, users are free to:
-
Share: Copy and redistribute the material in any medium or format.
-
Adapt: Remix, transform, and build upon the material for any purpose, including commercial purposes.
These permissions are subject to the following conditions:
-
Attribution: Users must give appropriate credit to the original author(s), retain relevant copyright notices, provide a link to the license, and indicate if changes were made. Attribution must not imply endorsement by the author(s), copyright holder, or journal.
-
ShareAlike: If users remix, transform, or build upon the material, they must distribute their contributions under the same CC BY-SA 4.0 license or a compatible license.
-
No Additional Restrictions: Users may not apply legal terms or technological measures that legally restrict others from exercising the rights granted by the license.
For more information about the CC BY-SA 4.0 license, please visit:
https://creativecommons.org/licenses/by-sa/4.0/
Third-Party Content
Third-party materials, including figures, tables, photographs, and illustrations, are covered by the article's CC BY-SA 4.0 license only when their inclusion under that license is authorized.
Materials excluded from the license must be clearly identified with appropriate copyright notices and attribution. Authors are responsible for obtaining any necessary permissions, while users must comply with the applicable licensing terms or permissions when reusing excluded materials.
Copyright and Licensing Information
Copyright and licensing information should be clearly displayed on each published article's landing page and included in the full-text versions (HTML and PDF), where applicable, to ensure transparency regarding copyright ownership and permitted reuse.
