Classification techniques can be used in microarray analysis to predict sample phenotypes based on gene expression patterns. While novel and microarray-specific classification tools are constantly being developed, the existing body of pattern recognition and prediction algorithms provide effective tools (35). Dudoit and colleagues (36) offer a practical comparison of methods for the classification of tumors using gene expression data. Relevant tools from the statistical modeling tradition include: discriminant analysis (37), including linear, logistic, and more flexible discrimination techniques; tree-based algorithms, such as classification and regression trees (CART) by Breiman et al. (38) and variants; generalized additive models (39); and neural networks (7, 40, 41). Appropriate versions of these methods can be used for both classification and prediction of quantitative responses such as continuous measures of disease aggressiveness. Some of these methods are briefly reviewed here.

0 0

Post a comment