作者: Andrew V. Kossenkov , Michael F. Ochs
DOI: 10.1016/S0076-6879(09)67003-8
关键词:
摘要: We explore a number of matrix factorization methods in terms their ability to identify signatures biological processes large gene expression study. focus on the these find ontology enhancement and interpretation samples. Two Bayesian approaches, Decomposition (BD) Factor Regression Modeling (BFRM), perform best. Differences strength between samples suggest that BD will be most useful for systems modeling BFRM biomarker discovery.