|
Abstract:
|
Survival prediction from high-dimensional genomic data is subject to much research effort in today's medicine. When dealing with these kinds of data, one has to take into account the problem of both censored observations and that the number of genetic variables p usually by far exceeds the number of patients n(p>>n). Most of the proposed methods handle this by combining the Cox proportional hazards model with some estimation regularization technique. We have investigated the performance of seven such methods; two subset selection methods, principal components regression (PCR), supervised PCR, LASSO, ridge regression, and our own modified partial least squares (PLS) Cox regression method. We will describe the estimation techniques and criteria used for evaluating the performance of the methods, and present the results from applications to three published microarray gene expression datasets.
|
