support vector machines (SVMs) in order to fit the proposed classifier. The structure of the paper is as follows. In “materials and methods,” we provide background on the data structures observed and the motivation based on biomarker combinations, which leads to the use of linear discriminant functions. We also provide a review of LASSO estimation (Tibshirani [11]) in this section. The latter two techniques are then involved in the proposed estimation procedure, described in “results and discussion.” There, we also describe how to implement the proposed method using software for SVMs. Issues of model selection are also discussed. We describe the application of the proposed methodologies to simulated data and data from a recent cancer profiling study (Dhanasekaran et al [3]) in “prostate cancer gene expression data.” Finally, some concluding remarks are made in “conclusion.”
