Multi-View Learning in Biomedical Applications in the Big Data Era
Roberto Tagliaferri, Angela Serra and Paola Galdi
Dipartimento di Scienze Aziendali - Management & Innovation Systems/DISA-MIS
Università di Salerno
In the era of big data, due to the richness and variety of available data sets, new horizons has been opened for investigations in the bio-medical field. Aim of these methodologies is the construction of an integrated base of knowledge derived from heterogeneous sources. Multi-view learning is the field of machine learning concerning the analysis of multi-modal data, i.e. patterns represented by different feature sets extracted from multiple data sources. Recently, multi-view learning methodologies have become increasingly popular for solving classification and clustering problems, and a high number of biomedical applications based on multi-view data has been recorded in literature.
For example, in bioinformatics, multi-view learning can be applied to multiple experiments investigating different facets of the same phenomena, such as gene expression, microRNA expression, protein-protein interactions, genome wide association and so on, in order to capture information regarding different aspects of biological systems. In the same way, in neurosciences big data mining can benefit from different imaging modalities that allow studying different features of the nervous system (e.g. structural vs. functional organization). Compared to the limited perspective offered by single-view analyses, the integration of multiple views can provide a deeper understanding of the underlying principles governing complex systems.
In this seminar, we review the existing multi-view methodologies to discuss their operation modes and principles, with the goal of increasing their further development in the bio-medical area.