Agenda

29 Set 2016 12:00

Kernel-⁠based learning methods applied to biological data

Campus Scientifico via Torino - edificio ZETA, Sala Riunioni

prof. Carlos Pereira, Polytechnic Institute of Coimbra

Abstract:
Kernel-based learning methods like support vector machines are among the best learning techniques for many benchmark datasets and real world applications. The advantages of the application of kernel machines to analyze big biological data are presented.

The main paradigms of learning from data are discussed, including neuro-fuzzy systems, genetic algorithms, and support vector machines.

Some recently developed research projects are described, including the problems of text mining from biomedical literature, protein function prediction and interaction.

In particular, a significant work has been done to solve the problem of protein classification but no ideal solution has been reached yet, mainly due to the complex nature and wide diversity of the field. A project aiming the development of efficient incremental kernel machines for biological data analysis, based on incremental kernel machines is presented.

A practical implementation, illustrating the example of protein classification is explained.

Lingua

L'evento si terrà in italiano

Organizzatore

Nicoletta Cocco

Cerca in agenda