Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Parametric quantile regression models for continuous proportions data : an evaluation of mean versus median modeling beyond beta
Mini Dissertation (MSc (Advanced Data Analytics))--University of Pretoria, 2025.
Saved in:
| Other Authors: | |
|---|---|
| Format: | Thesis |
| Language: | en_US |
| Published: |
University of Pretoria
2025
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613450372382720 |
|---|---|
| access_status_str | Open Access |
| author2 | Burger, Divan A. |
| author_browse | Burger, Divan A. |
| author_facet | Burger, Divan A. |
| collection | Thesis |
| dc_rights_str_mv | © 2023 University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria. |
| description | Mini Dissertation (MSc (Advanced Data Analytics))--University of Pretoria, 2025. |
| format | Thesis |
| id | oai:repository.up.ac.za:2263/100786 |
| institution | University of Pretoria (South Africa) |
| language | en_US |
| last_indexed | 2026-06-10T12:36:20.438Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from UPSpace — University of Pretoria Institutional Repository |
| publishDate | 2025 |
| publishDateRange | 2025 |
| publishDateSort | 2025 |
| publisher | University of Pretoria |
| publisherStr | University of Pretoria |
| record_format | dspace |
| source_str | UPSpace — University of Pretoria Institutional Repository |
| spelling | oai:repository.up.ac.za:2263/100786 Parametric quantile regression models for continuous proportions data : an evaluation of mean versus median modeling beyond beta Burger, Divan A. u16001223@tuks.co.za Weideman, Maricelle UCTD Sustainable Development Goals (SDGs) Bounded data Outliers Robustness Median regression Mean-based models Parametric models Mini Dissertation (MSc (Advanced Data Analytics))--University of Pretoria, 2025. In the modeling of bounded data, outliers or influential values present unique challenges, particularly when dealing with continuous proportions data. Traditional models, such as the beta regression model, although widely adopted, lack robustness against outliers, thus motivating the need for alternative models capable of addressing these limitations. This dissertation provides a comprehensive evaluation of various parametric models, including the beta, beta rectangular, Kumaraswamy, and Johnson-t models, emphasizing their robustness in handling outliers. A simulation study was conducted to examine the performance of each model under scenarios with and without outliers, measuring bias and coverage for key parameters. Results confirm the beta regression model’s sensitivity to outliers, as evidenced by increased bias and reduced coverage when influential values were introduced. In contrast, the Johnson-t regression model maintained stability in both bias and coverage, demonstrating greater resilience in outlier-inclusive datasets. Application to the Australian Institute of Sport data set further validated these findings, highlighting the Johnson-t model’s effectiveness in achieving robust median regression compared to mean-based approaches, which were less reliable with outliers. This study concludes that while beta regression remains popular for bounded data, the Johnson-t regression model offers a preferable alternative due to its robustness in median modeling, a critical factor in data analysis where influential values cannot be ignored. Statistics MSc (Advanced Data Analytics) Unrestricted Faculty of Natural and Agricultural Sciences SDG-09: Industry, innovation and infrastructure 2025-02-12T12:50:40Z 2025-02-12T12:50:40Z 2025-04 2025-02 Mini Dissertation * A2025 http://hdl.handle.net/2263/100786 https://cran.r-project.org/web/packages/DAAG/DAAG.pdf en_US © 2023 University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria. application/pdf University of Pretoria |
| spellingShingle | UCTD Sustainable Development Goals (SDGs) Bounded data Outliers Robustness Median regression Mean-based models Parametric models Parametric quantile regression models for continuous proportions data : an evaluation of mean versus median modeling beyond beta |
| title | Parametric quantile regression models for continuous proportions data : an evaluation of mean versus median modeling beyond beta |
| title_full | Parametric quantile regression models for continuous proportions data : an evaluation of mean versus median modeling beyond beta |
| title_fullStr | Parametric quantile regression models for continuous proportions data : an evaluation of mean versus median modeling beyond beta |
| title_full_unstemmed | Parametric quantile regression models for continuous proportions data : an evaluation of mean versus median modeling beyond beta |
| title_short | Parametric quantile regression models for continuous proportions data : an evaluation of mean versus median modeling beyond beta |
| title_sort | parametric quantile regression models for continuous proportions data an evaluation of mean versus median modeling beyond beta |
| topic | UCTD Sustainable Development Goals (SDGs) Bounded data Outliers Robustness Median regression Mean-based models Parametric models |
| url | http://hdl.handle.net/2263/100786 https://cran.r-project.org/web/packages/DAAG/DAAG.pdf |