2009 | OriginalPaper | Chapter
Bayesian Dimension Reduction Models for Microarray Data
Author : Albert D. Shieh
Published in: Adaptive and Natural Computing Algorithms
Publisher: Springer Berlin Heidelberg
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
High dimensionality, missing values, noise, and outliers are standard problems in gene expression data and are usually dealt with separately. In this paper, we propose an ideal point model that performs feature extraction, imputes missing values, and is robust to noise and outliers in a unified and unsupervised framework. We use the simplifying assumption that genes are either expressed or not expressed in order to obtain a parsimonious model. We present a fast Bayesian method for estimating the large number of parameters in the ideal point model. We apply the ideal point model to a leukemia data set, where it outperforms independent component analysis (ICA), a state of the art unsupervised feature extraction method.