Explore multiblock data analysis techniques in this 40-minute conference talk by Kristian Hovde Liland from Norwegian University of Life Sciences. Learn about the fundamentals of multiblock data fusion, including shared sample modes, linking blocks, and scaling. Discover standard methods like simultaneous component analysis, SOPLS, and ROSA. Gain insights into interpreting findings, handling heterogeneous data, and identifying local and distinct components. Apply these concepts to various fields such as omics data linking, spectroscopy-based characterization and prediction, and sensory analysis. Get a preview of the upcoming Wiley book "Multiblock Data Fusion in Statistics and Machine Learning – Applications in the Natural and Life Sciences" by Smilde, Næs, and Liland.
Overview
Syllabus
Introduction
Multiblock data
Shared sample mode
Linking blocks
Scaling
Motivation
Interpreting findings
Simultaneous component analysis
Heterogeneous data
Local and distinct components
Concatenate version
Typical data
SOPLS
ROSA
Bonus
Resources
Thanks
Taught by
Chemometrics & Machine Learning in Copenhagen
