Multivariate techniques are very intuitive methods, but the mathematics behind are sometimes complex. At CoDaS Lab, we enjoy studying matrix algebra problems and machine learning algorithms that deal with clear practical applications. Often, those practical applications come from our collaboration with other researchers or practitioners (biologists, ecologist, medical doctors, network engineers, etc.), and our role is to provide a technical solution for them.
Topics of interest at CoDaS Lab include (but not limited to) the selection and validation of multivariate models, the development of statistical inference and sample size strategies, the connection between data and covariance/distance matrices, the derivation of computational solutions for Big Data, and the development of new exploratory data analysis techniques, notably of constrained (e.g., sparse) multivariate models for omics data: genomics, epigenomics, metagenomics, metabolomics, etc.
