Granular Computing Techniques for Bioinformatics Pattern Recognition Problems in Non-metric Spaces

Abstract

Computational intelligence and pattern recognition techniques are gaining more and more attention as the main computing tools in bioinformatics applications. This is due to the fact that biology by definition, deals with complex systems and that computational intelligence can be considered as an effective approach when facing the general problem of complex systems modelling. Moreover, most data available on shared databases are represented by sequences and graphs, thus demanding the definition of meaningful dissimilarity measures between patterns, which are often non-metric in nature. Especially in such cases, evolutive and fully automatic machine learning systems are mandatory for dealing with parametric dissimilarity measures and/or for performing suitable feature selection. Besides other approaches, such as kernel methods and embedding in dissimilarity spaces, granular computing is a very promising framework not only for designing effective data-driven modelling systems able to determine automatically the correct representation (abstraction) level, but also for giving to field-experts (biologists) the possibility to investigate information granules (frequent substructures) that have been discovered by the machine learning system as the most relevant for the problem at hand. We expect that many important discoveries in biology and medicine in the next future will be determined by an increasingly stronger integration between the ongoing research efforts of natural sciences and modern inductive modelling tools based on computational intelligence, pattern recognition and granular computing techniques.