Employing Adaptive Bridge with Sliced Inverse Regression Methods to Identify Factors Influencing the Severity of Lung Cancer Infection

Abstract

This paper proposes the Adaptive Bridge–Sliced Inverse Regression (SIR) method for simultaneous variable selection and dimension reduction in high-dimensional biomedical data. The approach integrates adaptive penalization with the SIR framework to achieve robust estimation and interpretability. Simulation experiments demonstrate that the proposed method consistently attains lower Mean Median Absolute Deviation (MMAD) and Standard Deviation (SD) values compared with SIR–LASSO and SIR–MCP. A real clinical application on lung cancer data confirms its practical efficiency in identifying the most influential factors affecting disease severity.

Keywords: Adaptive Bridge; Sliced Inverse Regression; Variable Selection; Dimension Reduction; Lung Cancer Severity.

2010 Mathematics Subject Classification. 26A25; 26A35