pdmclass

pdmclass applies penalized regression methods to classify microarray cancer samples and rank genes by their contribution to molecular classification.

Key Features:

• Penalized Regression Methods: Implements partial least squares (PLS), principal components regression (PCR), and ridge regression for analysis of microarray data.
• Optimal Scoring Adaptation: Adapts PLS, PCR, and ridge regression for classification using the optimal scoring algorithm.
• Gene Ranking Procedure: Ranks genes based on fitted regression models to indicate gene significance in the classification models.

Scientific Applications:

• Cancer Classification: Produces molecular classifications of tumors from microarray data to distinguish between cancer types.
• Application to Microarray Studies: The implemented methodologies have been applied to two microarray studies in cancer.

Methodology:

Uses regularized regression models—principal components regression (PCR), partial least squares (PLS), and ridge regression—adapted for classification via the optimal scoring algorithm, with a gene ranking procedure derived from the fitted models.