Overall model fit criteria for multivariate integrative models

Supervisors: Kim-Anh Lê Cao

Available for: PhD/MSc

Location: Melbourne Integrative Genomics, University of Melbourne

Project title: Overall model fit criteria for multivariate integrative models

Description: Sparse Partial Least Squares regression (sPLS) is a useful multivariate method to integrate two biological datasets of heterogeneous sources such as ’omics data. This dimension reduction technique includes LASSO penalties to select the useful biological features that are correlated across datasets. However, appropriate statistical criteria in this complex multivariate setting have not yet been developed to assess the overall model fit.

The project will investigate several measures for sPLS and for other matrix factorization methods developed in our lab.