A1064
Title: An efficient approach for identifying important biomarkers for biomedical diagnosis
Authors: Jing-Wen Huang - Academia Sinica (Taiwan) [presenting]
Frederick Kin Hing Phoa - Academia Sinica (Taiwan)
Yan-Hong Chen - Academia Sinica (Taiwan)
Yan-Han Lin - National Taiwan University (Taiwan)
Shau Ping Lin - National Taiwan University (Taiwan)
Abstract: The purpose is to explore the challenges associated with biomarker identification for diagnostic purposes in biomedical experiments and propose a novel approach inspired by the analysis of supersaturated designs to handle the above challenging scenario via the generalization of the Dantzig selector. To improve the efficiency of the regularization method, a transformation from an inherent nonlinear programming problem with a nonlinear link function is introduced into a linear programming framework under a reasonable assumption on the logistic probability range. The use of the method is illustrated in an experiment with a binary response, showing superior performance in biomarker identification studies compared to conventional analysis. The proposed method does not merely serve as a variable/biomarker selection tool; its ranking of variable importance provides valuable reference information for practitioners to reach informed decisions regarding the prioritization of factors for further investigations.
