Heterogeneous Component Analysis.pdfVIP

Heterogeneous Component Analysis Shigeyuki Oba 1 , Motoaki Kawanabe 2 , Kraus Robert Mu?ller 2,3 , and Shin Ishii 1,4 1. Nara Institute of Science and Technology, Japan 2. Fraunhofer FIRST.IDA, Germany 3. Department of Computer Science Technical University Berlin, Germany 4. Kyoto University, Japan shige-o@is.naist.jp Abstract In bioinformatics it is often desirable to combine data from various measurement sources and thus structured feature vectors are to be analyzed that possess different intrinsic blocking characteristics (e.g., different patterns of missing values, obser- vation noise levels, effective intrinsic dimensionalities). We propose a new ma- chine learning tool, heterogeneous component analysis (HCA), for feature extrac- tion in order to better understand the factors that underlie such complex structured heterogeneous data. HCA is a linear block-wise sparse Bayesian PCA based not only on a probabilistic model with block-wise residual variance terms but also on a Bayesian treatment of a block-wise sparse factor-loading matrix. We study vari- ous algorithms that implement our HCA concept extracting sparse heterogeneous structure by obtaining common components for the blocks and specific compo- nents within each block. Simulations on toy and bioinformatics data underline the usefulness of the proposed structured matrix factorization concept. 1 Introduction Microarray and other high-throughput measurement devices have been applied to examine speci- mens such as cancer tissues of biological and/or clinical interest. The next step is to go towards combinatorial studies in which tissues measured by two or more of such devices are simultaneously analyzed. However, such combinatorial studies inevitably suffer from differences in experimental conditions, or, even more complex, from different measurement technologies. Also, when concate- nating a data set from different measurement sources, we often observe systematic missing parts in a dataset (e.g., Fig 3A). Mo