Properties and real data application of the Sine Alpha Power Burr XII regression model

Abstract

We propose the Sine Alpha Power Burr XII (SAP Burr XII) distribution, obtained by applying the Sine Alpha Power transformation to the Burr XII distribution. and develop an accelerated failure time (AFT) regression model, with parameters estimated via maximum likelihood. A simulation study across various sample sizes confirms the consistency of the estimators. Additionally, model diagnostics based on Cox–Snell, deviance, and randomized quantile residuals are examined. Application to a real lung cancer dataset demonstrates that the SAP Burr XII model outperforms classical AFT models.