Weighted accelerated failure time model

Nayaka Bandaralage, Ayesha Madhushani Rathnayake (2025) Weighted accelerated failure time model. Masters thesis, Memorial University of Newfoundland.

[English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (707kB)

Abstract

The accelerated failure time (AFT) model is widely used in survival analysis and auxiliary information can be used to improve the efficiency of the model. We developed a weighted AFT model by using empirical likelihood probabilities as weights based on information from previous studies. The proposed model effectively overcomes the challenges associated with managing censored observations, resulting in more reliable and accurate estimates. Theoretical justifications of the proposed model are developed. A comprehensive simulation study was conducted to assess the effectiveness of the proposed weighted models, incorporating both partial and complete auxiliary information. Both the Standard Accelerated Failure Time (AFT) and AFT with Generalized Estimating Equations (AFTGEE) models were employed for this comparative analysis. The simulation results suggest that when estimating coefficients, weighted models incorporating complete or partial auxiliary information on the linked covariate provide more accurate estimates compared to the model without any weights. Finally, the proposed method was implemented on a real dataset, illustrating its ability to accurately determine coefficients, minimize standard errors, and enhance significance levels by incorporating auxiliary information.

Actions (login required)

View Item